docs(readme): clarify /loop runs a prompt or custom command, not built-ins

- stopLoop re-arms the poll ticker after removing one loop, so stopping one of
  several no longer strands the rest with no tick scheduled (jatmn).
- a bare self-paced /command is rejected with a pointer to the interval form:
  self-paced drives a free-form goal via the playbook + LOOP: control line, which
  a custom command's expanded template never carries, so it would silently
  degrade to adaptive polling (jatmn).
- /resume only tears loops down on an actual session-id change; /resume latest or
  /resume <currentID> can resolve to the already-active session, whose loops are
  its own, not a "previous" session's (CodeRabbit).
- tests: skipped-behind-modal loop stays scheduled, footer carries the next-wake
  time, the confirmed /clear actually wipes the transcript, plus multi-loop stop
  keeping the ticker alive and self-paced-/command rejection.

.github/workflows/ci.yml

@@ -100,9 +100,10 @@ jobs:
           go-version-file: go.mod
           cache: true
 
-      # govulncheck and deadcode resolve a toolchain from their own modules, which
-      # can drop below the version go.mod requires and then fail to load our
-      # packages. Pin GOTOOLCHAIN to the go.mod toolchain so both run under it.
+      # govulncheck, deadcode, and golangci-lint each resolve a toolchain from
+      # their own modules, which can drop below the version go.mod requires and
+      # then fail to load our packages. Pin GOTOOLCHAIN to the go.mod toolchain
+      # so all three run under it.
       - name: Pin toolchain from go.mod
         run: |
           toolchain="$(awk '/^toolchain /{print $2}' go.mod)"
@@ -120,3 +121,13 @@ jobs:
       - name: deadcode (advisory)
         continue-on-error: true
         run: go run golang.org/x/tools/cmd/[email protected] -test=false ./...
+
+      # Advisory: catches what deadcode's whole-program reachability analysis
+      # doesn't, unused private functions/assignments reachable within a
+      # package but never actually called, plus staticcheck-style correctness
+      # and readability issues. Scoped to a few linters rather than the full
+      # default battery, and non-blocking, while the existing findings across
+      # the repo are cleaned up incrementally (see #527).
+      - name: golangci-lint (advisory)
+        continue-on-error: true
+        run: go run github.com/golangci/golangci-lint/v2/cmd/[email protected] run --enable-only unused,ineffassign,staticcheck ./...

.gitignore

@@ -37,8 +37,13 @@ report.[0-9]_.[0-9]_.[0-9]_.[0-9]_.json
 
 /zero
 /zero.exe
+/zero-skills
 /snake-game.html
 
 # Superpowers skill artifacts (brainstorm mockups, specs, plans) — local only
 .superpowers/
 docs/superpowers/
+
+# Termux/Android build artifacts
+zero-termux-arm64
+zero-linux-sandbox

AGENTS.md

@@ -47,6 +47,12 @@ Tips:
 - In a monorepo, drop a narrower `AGENTS.md` in each sub-tree (e.g. `services/api/AGENTS.md`). Zero picks those up automatically when you launch from inside the sub-tree.
 - A YAML frontmatter block (`---\n...\n---`) at the top is preserved verbatim in the injected prompt but is not parsed for `globs:` or `alwaysApply:` scoping today — keep the body self-contained.
 
+### Personal guidelines, across every project
+
+For preferences that follow *you*, not a specific repo (tone, tooling habits, workflow), drop a `ZERO.md` in your user config directory: `~/.config/zero/ZERO.md` on Linux/macOS, `%AppData%\Roaming\zero\ZERO.md` on Windows — the same directory as `config.json` and your personal specialists. Same format and 8 KiB cap as the project files above, and the same case-insensitive basename match.
+
+This file is injected as its own `## User guidelines` section, before the project's `AGENTS.md`/`ZERO.md`, and is labeled as personal preference in the prompt: project guidelines are the later, more specific instruction and take precedence over it when the two conflict.
+
 ## 2. Custom specialists
 
 Specialists are Zero's sub-agents. Three scopes, in priority order:
@@ -307,6 +313,7 @@ A team that wants every contributor's Zero to behave the same way commits:
 Each contributor adds only:
 
 - `~/.config/zero/config.json` — their personal API keys, theme, default mode.
+- `~/.config/zero/ZERO.md` — personal preferences that follow them across every project (see section 1).
 - `~/.local/share/zero/skills/` — personal skills they keep across projects.
 
 That's it. Run `zero` from the repo root and the agent has the team's full instruction set, every contributor's personal setup, and nothing else.

README.md

@@ -186,6 +186,7 @@ Common slash commands:
 | `/spec`, `/plan` | draft and review a plan before building |
 | `/image` | attach an image for vision-capable models |
 | `/resume`, `/rewind` | continue or roll back local sessions |
+| `/loop` | repeat a prompt or custom `/command` on an interval (`/loop 5m /babysit-prs`) or self-paced |
 | `/compact`, `/context` | manage context usage |
 | `/permissions`, `/tools` | inspect available tools and policy |
 | `/add-dir` | allow an extra write directory for this session |
@@ -283,6 +284,39 @@ zero cron             scheduled agent jobs
 zero update           check for newer releases
 ```
 
+## Extending Zero
+
+### Project and personal instructions
+
+Zero appends project-specific guidance to the system prompt from the first
+`AGENTS.md`, `ZERO.md`, or `.zero/AGENTS.md` file found in each directory from
+the git root down to your current working directory (checked in that order
+per directory). Files are injected general-to-specific, capped at 8 KiB per
+file and 32 KiB total.
+
+A personal `ZERO.md` under `config.UserConfigDir()/zero/ZERO.md`
+(`$XDG_CONFIG_HOME/zero/ZERO.md` or `~/.config/zero/ZERO.md` on Linux/macOS,
+`%AppData%\Roaming\zero\ZERO.md` on Windows) applies across every workspace, ahead of any project guidelines.
+
+### Plugins
+
+Plugins are discovered from `~/.config/zero/plugins/<name>/plugin.json` (user
+scope — `$XDG_CONFIG_HOME` or `~/.config` on every OS, independent of the
+`config.UserConfigDir()` path used above) and `<cwd>/.zero/plugins/<name>/plugin.json`
+(project scope — resolved from the current working directory, not the repo
+root), and managed with `zero plugins`. A manifest can declare:
+
+- `tools` — custom tools (`command`, `args`, `inputSchema`, and a
+  `permission` of `prompt` or `deny`; `allow` is honored only when manifest tool
+  auto-approval is enabled)
+- `hooks` — commands run on `beforeTool`, `afterTool`, `sessionStart`, or
+  `sessionEnd`
+- `prompts` and `skills` — additional prompt/skill files
+
+MCP servers (`zero mcp`) and standalone markdown skills (`zero skills`) use
+the same extension points and can also be wired up outside of a plugin
+manifest.
+
 ## Appearance And Accessibility
 
 | Control | Effect |

action.yml

@@ -180,6 +180,7 @@ runs:
         ZERO_INPUT_ADD_DIR: ${{ inputs.add-dir }}
         ZERO_INPUT_WORKTREE: ${{ inputs.worktree }}
         ZERO_INPUT_OUTPUT_FORMAT: ${{ inputs.output-format }}
+        ACTION_PATH: ${{ github.action_path }}
         RUNNER_TEMP: ${{ runner.temp }}
       run: |
         set -euo pipefail
@@ -247,10 +248,9 @@ runs:
         echo "exit-code=${code}" >> "${GITHUB_OUTPUT}"
         echo "output-file=${output_file}" >> "${GITHUB_OUTPUT}"
 
-        # Best-effort one-line summary: last non-empty line of the captured output.
-        summary="$(grep -v '^[[:space:]]*$' "${output_file}" | tail -n 1 || true)"
-        # Keep the summary single-line and bounded for the step output.
-        summary="$(printf '%s' "${summary}" | cut -c1-280)"
+        # Best-effort one-line summary. Structured output ends with lifecycle
+        # metadata, so parse final/error events instead of tailing stdout.
+        summary="$(node "${ACTION_PATH}/scripts/action-summary.mjs" "${ZERO_INPUT_OUTPUT_FORMAT}" "${output_file}" || true)"
         {
           echo "summary<<ZERO_EOF"
           printf '%s\n' "${summary}"

docs/INSTALL.md

@@ -173,6 +173,61 @@ Linux native sandboxing also requires Bubblewrap to be installed.
 
 macOS uses the system sandbox and does not need an extra helper binary.
 
+### Termux (Android)
+
+Zero can run natively on Android via [Termux](https://termux.dev/). Build with
+`GOOS=android` to avoid the `faccessat2` syscall that is blocked by Samsung's
+seccomp filter on Android:
+
+```bash
+# Install Go in Termux
+pkg install golang
+
+# Build Zero for Android
+git clone https://github.com/Gitlawb/zero.git
+cd zero
+CGO_ENABLED=0 GOOS=android GOARCH=arm64 go build -ldflags="-s -w" -o zero ./cmd/zero
+
+# Move into PATH
+mv zero ~/.local/bin/
+```
+
+> **Why `GOOS=android`?** Go 1.26+ detects `runtime.GOOS == "android"` and skips
+> the `faccessat2` syscall inside `os/exec.findExecutable`, falling back to
+> permission-bit checks. Without this flag, Android's seccomp sends SIGSYS and
+> kills the process whenever Zero looks up a binary on `PATH` (git, sh, etc.).
+
+**DNS.** Android does not expose `/etc/resolv.conf`. Go's pure-Go DNS resolver
+needs one. Use `proot` to bind-mount Termux's resolver config:
+
+```bash
+pkg install proot
+proot -b "$PREFIX/etc/resolv.conf:/etc/resolv.conf" zero
+```
+
+Create a wrapper at `~/.local/bin/zero` to avoid typing proot every time:
+
+```bash
+#!/data/data/com.termux/files/usr/bin/bash
+exec proot -b "$PREFIX/etc/resolv.conf:/etc/resolv.conf" ~/.local/bin/zero.bin "$@"
+```
+
+**Scroll.** On native Termux (not under PRoot), mouse scrolling works out of the
+box. The TUI uses Bubble Tea's `AllMotion` mouse mode by default. If you run Zero
+inside PRoot (e.g. through proot-distro), the scroll fix activates `CellMotion`
+to avoid PRoot's ptrace interference with the 1003 escape sequence.
+
+**Providers.** Zero works with any OpenAI-compatible provider on Termux. For
+example, to use OpenCode Zen's free tier:
+
+```bash
+zero providers add opencode \
+  --name opencode \
+  --model deepseek-v4-flash-free \
+  --base-url https://opencode.ai/zen/v1 \
+  --set-active
+```
+
 Windows source builds can use the main `zero.exe` as the command runner and setup
 helper through Zero's built-in self-dispatch path. If you want a release-style
 layout anyway, build the standalone helper executables next to `zero.exe`:

docs/STREAM_JSON_PROTOCOL.md

@@ -41,6 +41,7 @@ events when present.
 
 ```json
 { "schemaVersion": 2, "type": "run_start", "runId": "run_20260603_abc123", "sessionId": "zero_20260603100000_abc123", "cwd": "/repo", "provider": "openai", "model": "gpt-4.1", "apiModel": "gpt-4.1" }
+{ "schemaVersion": 2, "type": "reasoning", "runId": "run_20260603_abc123", "delta": "Thinking..." }
 { "schemaVersion": 2, "type": "text", "runId": "run_20260603_abc123", "delta": "..." }
 { "schemaVersion": 2, "type": "tool_call", "runId": "run_20260603_abc123", "id": "call_1", "name": "read_file", "args": { "path": "README.md" }, "sideEffect": "read" }
 { "schemaVersion": 2, "type": "permission_request", "runId": "run_20260603_abc123", "id": "call_2", "name": "write_file", "action": "prompt", "permission": "prompt", "permissionMode": "ask", "sideEffect": "write", "reason": "Creates or overwrites files." }
@@ -51,6 +52,10 @@ events when present.
 { "schemaVersion": 2, "type": "run_end", "runId": "run_20260603_abc123", "status": "success", "exitCode": 0 }
 ```
 
+`reasoning` events carry live model reasoning/status deltas for providers that
+stream them separately from answer text. They are liveness/progress events only:
+they are not folded into `text` or the final answer.
+
 Permission events may include structured sandbox metadata:
 
 ```json

internal/agent/async_diagnostics.go

@@ -0,0 +1,185 @@
+package agent
+
+import (
+	"context"
+	"path/filepath"
+	"sort"
+	"strings"
+	"sync"
+	"time"
+)
+
+// Post-edit LSP diagnostics used to run synchronously inside edit_file /
+// write_file: every edit blocked its tool result on a fresh-publish wait with
+// a quiet-period debounce (≥300ms floor that reset on each publish, 10s cap) —
+// the largest recurring latency tax in an edit-heavy session. Collection is
+// now asynchronous: a file is enqueued the moment a mutating tool finishes, a
+// single worker checks it while the rest of the tool batch (and any
+// self-correct verification) runs, and the loop drains completed results just
+// before building the NEXT provider request, appending errors as a user-role
+// nudge. The model always issues another request after an edit turn — it has
+// to read its tool results — so it still sees introduced errors at the same
+// decision point as before, without any tool call stalling on the language
+// server.
+//
+// A file re-edited before its check runs is checked once in its final state
+// (the checker re-reads the file); a file re-edited after its check completed
+// is re-enqueued and the newer result wins. When the worker is still busy at
+// drain time, the drain gives up after a short wait and the results are
+// delivered on the following turn instead — they stay accurate because any
+// further edit re-enqueues the file.
+
+// asyncDiagnosticsDrainTimeout bounds how long a turn waits on the in-flight
+// check before deferring delivery to the next turn. Most checks finish well
+// under it (debounce floor + analysis); a cold language server on a large
+// repo does not get to stall the loop the way the old inline 10s cap could.
+// A var so tests can shorten the wait.
+var asyncDiagnosticsDrainTimeout = 3 * time.Second
+
+// asyncDiagnosticsFinalDrainTimeout is the wait used when the run is about to
+// FINALIZE (natural completion or the max-turns summary): there is no later
+// turn to defer to, so the gate waits out the old inline-era budget rather
+// than silently dropping an error the last edit introduced. A var so tests
+// can shorten the wait.
+var asyncDiagnosticsFinalDrainTimeout = 10 * time.Second
+
+// asyncDiagnosticsNudge prefixes the drained diagnostics blocks; phrased like
+// the old inline block so the model reacts the same way.
+const asyncDiagnosticsNudge = "Diagnostics after your recent edits (fix any errors you introduced):\n"
+
+type asyncDiagnostics struct {
+	check         func(context.Context, string) string
+	workspaceRoot string
+
+	mu      sync.Mutex
+	queue   []string
+	queued  map[string]bool
+	results map[string]string
+	// working is non-nil while the worker goroutine runs and is closed when it
+	// exits, so drain can wait for quiescence without polling.
+	working chan struct{}
+}
+
+// newAsyncDiagnostics returns nil when check is nil, and every method on a nil
+// receiver is a no-op — callers wire it unconditionally.
+func newAsyncDiagnostics(check func(context.Context, string) string, workspaceRoot string) *asyncDiagnostics {
+	if check == nil {
+		return nil
+	}
+	return &asyncDiagnostics{
+		check:         check,
+		workspaceRoot: workspaceRoot,
+		queued:        map[string]bool{},
+		results:       map[string]string{},
+	}
+}
+
+// enqueue schedules changed files for a background check. Paths are
+// workspace-relative from Result.ChangedFiles (absolute when under a granted
+// extra write root), mirroring SelfCorrector's resolution.
+func (diagnostics *asyncDiagnostics) enqueue(ctx context.Context, files []string) {
+	if diagnostics == nil || len(files) == 0 {
+		return
+	}
+	diagnostics.mu.Lock()
+	defer diagnostics.mu.Unlock()
+	for _, file := range files {
+		path := diagnostics.absPath(file)
+		if diagnostics.queued[path] {
+			continue
+		}
+		diagnostics.queued[path] = true
+		// A completed result for this file describes a pre-edit state now.
+		delete(diagnostics.results, path)
+		diagnostics.queue = append(diagnostics.queue, path)
+	}
+	if diagnostics.working == nil && len(diagnostics.queue) > 0 {
+		done := make(chan struct{})
+		diagnostics.working = done
+		go diagnostics.work(ctx, done)
+	}
+}
+
+func (diagnostics *asyncDiagnostics) work(ctx context.Context, done chan struct{}) {
+	defer close(done)
+	for {
+		diagnostics.mu.Lock()
+		if len(diagnostics.queue) == 0 || ctx.Err() != nil {
+			diagnostics.working = nil
+			diagnostics.mu.Unlock()
+			return
+		}
+		path := diagnostics.queue[0]
+		diagnostics.queue = diagnostics.queue[1:]
+		// Un-mark before checking so an edit that lands DURING the check
+		// re-enqueues the file and it is checked again in its final state.
+		delete(diagnostics.queued, path)
+		diagnostics.mu.Unlock()
+
+		block := diagnostics.check(ctx, path)
+
+		diagnostics.mu.Lock()
+		if block != "" {
+			diagnostics.results[path] = block
+		} else {
+			delete(diagnostics.results, path)
+		}
+		diagnostics.mu.Unlock()
+	}
+}
+
+// drain waits briefly for the worker to go quiet, then formats and consumes
+// all completed error blocks as one nudge. Returns "" when there is nothing
+// to report or the worker is still busy (results then surface next turn).
+func (diagnostics *asyncDiagnostics) drain(ctx context.Context) string {
+	return diagnostics.drainWithin(ctx, asyncDiagnosticsDrainTimeout)
+}
+
+// drainFinal is drain with the finalization budget: called before the run's
+// last model request, where an undelivered result would otherwise be lost.
+func (diagnostics *asyncDiagnostics) drainFinal(ctx context.Context) string {
+	return diagnostics.drainWithin(ctx, asyncDiagnosticsFinalDrainTimeout)
+}
+
+func (diagnostics *asyncDiagnostics) drainWithin(ctx context.Context, timeout time.Duration) string {
+	if diagnostics == nil {
+		return ""
+	}
+	diagnostics.mu.Lock()
+	busy := diagnostics.working
+	diagnostics.mu.Unlock()
+	if busy != nil {
+		timer := time.NewTimer(timeout)
+		defer timer.Stop()
+		select {
+		case <-busy:
+		case <-timer.C:
+			return ""
+		case <-ctx.Done():
+			return ""
+		}
+	}
+	diagnostics.mu.Lock()
+	defer diagnostics.mu.Unlock()
+	if len(diagnostics.results) == 0 {
+		return ""
+	}
+	paths := make([]string, 0, len(diagnostics.results))
+	for path := range diagnostics.results {
+		paths = append(paths, path)
+	}
+	sort.Strings(paths)
+	blocks := make([]string, 0, len(paths))
+	for _, path := range paths {
+		blocks = append(blocks, diagnostics.results[path])
+	}
+	diagnostics.results = map[string]string{}
+	return asyncDiagnosticsNudge + strings.Join(blocks, "\n")
+}
+
+func (diagnostics *asyncDiagnostics) absPath(path string) string {
+	if filepath.IsAbs(path) {
+		return path
+	}
+	return filepath.Join(diagnostics.workspaceRoot, path)
+}

internal/agent/async_diagnostics_test.go

@@ -0,0 +1,338 @@
+package agent
+
+import (
+	"context"
+	"path/filepath"
+	"strings"
+	"sync"
+	"testing"
+	"time"
+
+	"github.com/Gitlawb/zero/internal/tools"
+	"github.com/Gitlawb/zero/internal/zeroruntime"
+)
+
+func TestAsyncDiagnosticsNilCollectorNoOps(t *testing.T) {
+	var diagnostics *asyncDiagnostics
+	if diagnostics != newAsyncDiagnostics(nil, "/ws") {
+		t.Fatal("nil check must yield a nil collector")
+	}
+	diagnostics.enqueue(context.Background(), []string{"a.go"})
+	if nudge := diagnostics.drain(context.Background()); nudge != "" {
+		t.Fatalf("nil collector drain = %q, want empty", nudge)
+	}
+}
+
+func TestAsyncDiagnosticsCollectsAndDrainsOnce(t *testing.T) {
+	// A real absolute root: a literal "/ws" is not filepath.IsAbs on Windows
+	// (no drive letter), which is exactly what the assertions below check.
+	root := t.TempDir()
+	var mu sync.Mutex
+	var checked []string
+	check := func(_ context.Context, absPath string) string {
+		mu.Lock()
+		checked = append(checked, absPath)
+		mu.Unlock()
+		if filepath.Base(absPath) == "clean.go" {
+			return ""
+		}
+		return "ERR " + filepath.Base(absPath)
+	}
+	diagnostics := newAsyncDiagnostics(check, root)
+	diagnostics.enqueue(context.Background(), []string{"broken.go", "clean.go"})
+
+	nudge := diagnostics.drain(context.Background())
+	if !strings.HasPrefix(nudge, asyncDiagnosticsNudge) {
+		t.Fatalf("nudge missing prefix: %q", nudge)
+	}
+	if !strings.Contains(nudge, "ERR broken.go") || strings.Contains(nudge, "clean.go") {
+		t.Fatalf("nudge = %q, want broken.go errors only", nudge)
+	}
+	mu.Lock()
+	for _, path := range checked {
+		if !filepath.IsAbs(path) || !strings.HasPrefix(path, root+string(filepath.Separator)) {
+			t.Fatalf("check received %q, want path under workspace root %q", path, root)
+		}
+	}
+	mu.Unlock()
+	if second := diagnostics.drain(context.Background()); second != "" {
+		t.Fatalf("second drain = %q, want empty (results are consumed)", second)
+	}
+}
+
+// A check still running at drain time defers delivery to the next drain
+// instead of blocking the turn.
+func TestAsyncDiagnosticsSlowCheckDefersToNextDrain(t *testing.T) {
+	previous := asyncDiagnosticsDrainTimeout
+	asyncDiagnosticsDrainTimeout = 20 * time.Millisecond
+	defer func() { asyncDiagnosticsDrainTimeout = previous }()
+
+	release := make(chan struct{})
+	check := func(_ context.Context, absPath string) string {
+		<-release
+		return "ERR " + filepath.Base(absPath)
+	}
+	diagnostics := newAsyncDiagnostics(check, "/ws")
+	diagnostics.enqueue(context.Background(), []string{"slow.go"})
+
+	start := time.Now()
+	if nudge := diagnostics.drain(context.Background()); nudge != "" {
+		t.Fatalf("drain during in-flight check = %q, want empty", nudge)
+	}
+	if elapsed := time.Since(start); elapsed > time.Second {
+		t.Fatalf("drain blocked %v, want bounded by the (shortened) timeout", elapsed)
+	}
+
+	close(release)
+	deadline := time.Now().Add(2 * time.Second)
+	for {
+		if nudge := diagnostics.drain(context.Background()); nudge != "" {
+			if !strings.Contains(nudge, "ERR slow.go") {
+				t.Fatalf("deferred nudge = %q", nudge)
+			}
+			return
+		}
+		if time.Now().After(deadline) {
+			t.Fatal("deferred diagnostics never delivered")
+		}
+		time.Sleep(5 * time.Millisecond)
+	}
+}
+
+// Re-enqueueing a file after its check completed replaces the stale result.
+func TestAsyncDiagnosticsReEditReplacesResult(t *testing.T) {
+	var mu sync.Mutex
+	response := "ERR old"
+	check := func(context.Context, string) string {
+		mu.Lock()
+		defer mu.Unlock()
+		return response
+	}
+	diagnostics := newAsyncDiagnostics(check, "/ws")
+
+	diagnostics.enqueue(context.Background(), []string{"a.go"})
+	waitForIdle(t, diagnostics)
+	mu.Lock()
+	response = "ERR new"
+	mu.Unlock()
+	diagnostics.enqueue(context.Background(), []string{"a.go"})
+
+	nudge := diagnostics.drain(context.Background())
+	if !strings.Contains(nudge, "ERR new") || strings.Contains(nudge, "ERR old") {
+		t.Fatalf("nudge = %q, want only the re-check result", nudge)
+	}
+}
+
+func waitForIdle(t *testing.T, diagnostics *asyncDiagnostics) {
+	t.Helper()
+	deadline := time.Now().Add(2 * time.Second)
+	for {
+		diagnostics.mu.Lock()
+		busy := diagnostics.working
+		diagnostics.mu.Unlock()
+		if busy == nil {
+			return
+		}
+		if time.Now().After(deadline) {
+			t.Fatal("worker never went idle")
+		}
+		time.Sleep(2 * time.Millisecond)
+	}
+}
+
+// changedFilesTool is a fake mutating tool reporting a changed file without
+// touching the filesystem.
+type changedFilesTool struct{}
+
+func (changedFilesTool) Name() string        { return "fake_edit" }
+func (changedFilesTool) Description() string { return "test mutating tool" }
+func (changedFilesTool) Parameters() tools.Schema {
+	return tools.Schema{Type: "object", Properties: map[string]tools.PropertySchema{}}
+}
+func (changedFilesTool) Safety() tools.Safety {
+	return tools.Safety{SideEffect: tools.SideEffectWrite, Permission: tools.PermissionAllow, Reason: "test"}
+}
+func (changedFilesTool) Run(context.Context, map[string]any) tools.Result {
+	return tools.Result{Status: tools.StatusOK, Output: "edited", ChangedFiles: []string{"main.go"}}
+}
+
+// End-to-end through Run: an edit turn does NOT carry diagnostics in the tool
+// result (the old inline path), but the next request contains the background
+// diagnostics nudge, before the model finalizes.
+func TestRunDeliversAsyncDiagnosticsNudgeNextTurn(t *testing.T) {
+	registry := tools.NewRegistry()
+	registry.Register(changedFilesTool{})
+	provider := &mockProvider{
+		turns: [][]zeroruntime.StreamEvent{
+			{
+				{Type: zeroruntime.StreamEventToolCallStart, ToolCallID: "call-1", ToolName: "fake_edit"},
+				{Type: zeroruntime.StreamEventToolCallDelta, ToolCallID: "call-1", ArgumentsFragment: `{}`},
+				{Type: zeroruntime.StreamEventToolCallEnd, ToolCallID: "call-1"},
+				{Type: zeroruntime.StreamEventDone},
+			},
+			{
+				{Type: zeroruntime.StreamEventText, Content: "done"},
+				{Type: zeroruntime.StreamEventDone},
+			},
+		},
+	}
+	root := t.TempDir()
+	var checkedPath string
+	var mu sync.Mutex
+
+	result, err := Run(context.Background(), "fix main.go", provider, Options{
+		Registry: registry,
+		Cwd:      root,
+		FileDiagnostics: func(_ context.Context, absPath string) string {
+			mu.Lock()
+			checkedPath = absPath
+			mu.Unlock()
+			return "main.go:1:1 error: boom"
+		},
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if result.FinalAnswer != "done" {
+		t.Fatalf("final answer = %q", result.FinalAnswer)
+	}
+	if len(provider.requests) != 2 {
+		t.Fatalf("expected 2 provider turns, got %d", len(provider.requests))
+	}
+
+	// The tool result itself must NOT block on / embed diagnostics.
+	for _, message := range provider.requests[1].Messages {
+		if message.Role == zeroruntime.MessageRoleTool && strings.Contains(message.Content, "boom") {
+			t.Fatalf("diagnostics leaked into the tool result: %q", message.Content)
+		}
+	}
+	// The nudge must be present in the second request as a user message.
+	found := false
+	for _, message := range provider.requests[1].Messages {
+		if message.Role == zeroruntime.MessageRoleUser && strings.HasPrefix(message.Content, asyncDiagnosticsNudge) {
+			if !strings.Contains(message.Content, "boom") {
+				t.Fatalf("nudge missing diagnostics: %q", message.Content)
+			}
+			found = true
+		}
+	}
+	if !found {
+		t.Fatal("no async diagnostics nudge in the follow-up request")
+	}
+	mu.Lock()
+	defer mu.Unlock()
+	if want := filepath.Join(root, "main.go"); checkedPath != want {
+		t.Fatalf("checked path = %q, want %q", checkedPath, want)
+	}
+}
+
+// When the run hits the maxTurns ceiling right after an edit turn, the
+// final-answer request must still carry the diagnostics nudge — otherwise an
+// error introduced by the last edit would go unreported in the summary.
+func TestRunDrainsAsyncDiagnosticsBeforeMaxTurnsFinalAnswer(t *testing.T) {
+	registry := tools.NewRegistry()
+	registry.Register(changedFilesTool{})
+	provider := &mockProvider{
+		turns: [][]zeroruntime.StreamEvent{
+			{
+				{Type: zeroruntime.StreamEventToolCallStart, ToolCallID: "call-1", ToolName: "fake_edit"},
+				{Type: zeroruntime.StreamEventToolCallDelta, ToolCallID: "call-1", ArgumentsFragment: `{}`},
+				{Type: zeroruntime.StreamEventToolCallEnd, ToolCallID: "call-1"},
+				{Type: zeroruntime.StreamEventDone},
+			},
+			{
+				{Type: zeroruntime.StreamEventText, Content: "summary"},
+				{Type: zeroruntime.StreamEventDone},
+			},
+		},
+	}
+
+	result, err := Run(context.Background(), "fix main.go", provider, Options{
+		Registry: registry,
+		Cwd:      t.TempDir(),
+		MaxTurns: 1,
+		FileDiagnostics: func(context.Context, string) string {
+			return "main.go:1:1 error: boom"
+		},
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if result.FinalAnswer != "summary" {
+		t.Fatalf("final answer = %q", result.FinalAnswer)
+	}
+	if len(provider.requests) != 2 {
+		t.Fatalf("expected the max-turns final-answer request, got %d requests", len(provider.requests))
+	}
+	found := false
+	for _, message := range provider.requests[1].Messages {
+		if message.Role == zeroruntime.MessageRoleUser && strings.HasPrefix(message.Content, asyncDiagnosticsNudge) {
+			found = true
+		}
+	}
+	if !found {
+		t.Fatal("max-turns final-answer request missing the diagnostics nudge")
+	}
+}
+
+// When the model finalizes while a slow check is still in flight (the
+// per-turn drain missed it), the finalization gate must wait it out and give
+// the model one more turn with the nudge instead of dropping the errors.
+func TestRunFinalizationGateDeliversLateDiagnostics(t *testing.T) {
+	previous := asyncDiagnosticsDrainTimeout
+	asyncDiagnosticsDrainTimeout = time.Millisecond
+	defer func() { asyncDiagnosticsDrainTimeout = previous }()
+
+	registry := tools.NewRegistry()
+	registry.Register(changedFilesTool{})
+	provider := &mockProvider{
+		turns: [][]zeroruntime.StreamEvent{
+			{
+				{Type: zeroruntime.StreamEventToolCallStart, ToolCallID: "call-1", ToolName: "fake_edit"},
+				{Type: zeroruntime.StreamEventToolCallDelta, ToolCallID: "call-1", ArgumentsFragment: `{}`},
+				{Type: zeroruntime.StreamEventToolCallEnd, ToolCallID: "call-1"},
+				{Type: zeroruntime.StreamEventDone},
+			},
+			{
+				{Type: zeroruntime.StreamEventText, Content: "done"},
+				{Type: zeroruntime.StreamEventDone},
+			},
+			{
+				{Type: zeroruntime.StreamEventText, Content: "done after fix"},
+				{Type: zeroruntime.StreamEventDone},
+			},
+		},
+	}
+
+	result, err := Run(context.Background(), "fix main.go", provider, Options{
+		Registry: registry,
+		Cwd:      t.TempDir(),
+		FileDiagnostics: func(context.Context, string) string {
+			// Slower than the (shortened) per-turn drain, well under the
+			// finalization budget.
+			time.Sleep(100 * time.Millisecond)
+			return "main.go:1:1 error: boom"
+		},
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if result.FinalAnswer != "done after fix" {
+		t.Fatalf("final answer = %q, want the post-nudge answer", result.FinalAnswer)
+	}
+	if len(provider.requests) != 3 {
+		t.Fatalf("expected a third request carrying the late nudge, got %d requests", len(provider.requests))
+	}
+	found := false
+	for _, message := range provider.requests[2].Messages {
+		if message.Role == zeroruntime.MessageRoleUser && strings.HasPrefix(message.Content, asyncDiagnosticsNudge) {
+			if !strings.Contains(message.Content, "boom") {
+				t.Fatalf("nudge missing diagnostics: %q", message.Content)
+			}
+			found = true
+		}
+	}
+	if !found {
+		t.Fatal("finalization gate did not deliver the late diagnostics nudge")
+	}
+}

internal/agent/command_prefix.go

@@ -2,6 +2,7 @@ package agent
 
 import (
 	"fmt"
+	"runtime"
 	"strings"
 
 	"github.com/Gitlawb/zero/internal/sandbox"
@@ -27,10 +28,20 @@ func proposedCommandPrefix(toolName string, args map[string]any) []string {
 		}
 		return nil
 	}
-	for _, tokens := range segments {
-		if !knownSafeCommandSegment(tokens) && sandbox.ValidCommandPrefix(tokens) {
-			return append([]string(nil), tokens...)
+	// Only propose approving a prefix of one segment when every other segment
+	// in the command is independently known-safe. Once a prefix is approved,
+	// shellExecutionArgsForApproval escalates the whole command (every
+	// segment, not just the approved one) to bypass the sandbox, so offering a
+	// prefix that leaves an MSYS-prone (or otherwise unsafe) segment uncovered
+	// would let that segment run unsandboxed without ever being reviewed.
+	for index, tokens := range segments {
+		if knownSafeCommandSegment(tokens) {
+			continue
 		}
+		if !sandbox.ValidCommandPrefix(tokens) || !otherSegmentsKnownSafe(segments, index) {
+			return nil
+		}
+		return append([]string(nil), tokens...)
 	}
 	if len(segments) == 0 || !sandbox.ValidCommandPrefix(segments[0]) {
 		return nil
@@ -38,6 +49,20 @@ func proposedCommandPrefix(toolName string, args map[string]any) []string {
 	return append([]string(nil), segments[0]...)
 }
 
+// otherSegmentsKnownSafe reports whether every segment other than the one at
+// skip is known-safe on its own.
+func otherSegmentsKnownSafe(segments [][]string, skip int) bool {
+	for index, tokens := range segments {
+		if index == skip {
+			continue
+		}
+		if !knownSafeCommandSegment(tokens) {
+			return false
+		}
+	}
+	return true
+}
+
 func matchCommandPrefix(toolName string, args map[string]any, options Options) (sandbox.CommandPrefixGrant, bool, bool) {
 	if !isShellCommandTool(toolName) || options.Sandbox == nil {
 		return sandbox.CommandPrefixGrant{}, false, false
@@ -168,10 +193,6 @@ func cleanPrefixRule(prefix []string) []string {
 	return cleaned
 }
 
-func safeRequestedPrefix(prefix []string, command []string) bool {
-	return safeRequestedPrefixForSegments(prefix, [][]string{command})
-}
-
 func safeRequestedPrefixForSegments(prefix []string, segments [][]string) bool {
 	if len(prefix) == 0 || !sandbox.ValidCommandPrefix(prefix) {
 		return false
@@ -196,14 +217,6 @@ func safeRequestedPrefixForSegments(prefix []string, segments [][]string) bool {
 	return matched
 }
 
-func safeShellCommandTokens(command string) ([]string, bool) {
-	segments, ok := safeShellCommandSegments(command)
-	if !ok || len(segments) != 1 {
-		return nil, false
-	}
-	return segments[0], true
-}
-
 func safeShellCommandSegments(command string) ([][]string, bool) {
 	command = strings.TrimSpace(command)
 	if command == "" {
@@ -275,7 +288,11 @@ func knownSafeCommandSegment(command []string) bool {
 	if len(command) == 0 {
 		return false
 	}
-	switch commandName(command[0]) {
+	name := commandName(command[0])
+	if runtime.GOOS == "windows" && tools.MsysProneCommandName(name) {
+		return false
+	}
+	switch name {
 	case "cat", "cd", "cut", "echo", "expr", "false", "grep", "head", "id",
 		"ls", "nl", "paste", "pwd", "rev", "seq", "stat", "tail", "tr",
 		"true", "uname", "uniq", "wc", "which", "whoami":

internal/agent/command_prefix_test.go

@@ -1,9 +1,11 @@
 package agent
 
 import (
+	"runtime"
 	"testing"
 
 	"github.com/Gitlawb/zero/internal/sandbox"
+	"github.com/Gitlawb/zero/internal/tools"
 )
 
 func TestProposedCommandPrefixUsesSafeSimpleCommands(t *testing.T) {
@@ -35,12 +37,37 @@ func TestProposedCommandPrefixHonorsValidatedRequestedPrefix(t *testing.T) {
 
 func TestProposedCommandPrefixSupportsSegmentedCommands(t *testing.T) {
 	got := proposedCommandPrefix("bash", map[string]any{"command": "ps aux | head -5"})
+	if runtime.GOOS == "windows" {
+		// head is MSYS-prone on Windows (#458), so proposedCommandPrefix must
+		// not offer "ps aux" as a reusable prefix here: approving it would
+		// escalate the whole command, including the uncovered head segment,
+		// to bypass the sandbox unreviewed. See
+		// TestProposedCommandPrefixRejectsPrefixLeavingUnsafeTailUncovered for
+		// the platform-independent regression coverage of this behavior.
+		if got != nil {
+			t.Fatalf("expected no prefix on Windows because head is MSYS-prone, got %#v", got)
+		}
+		return
+	}
 	want := []string{"ps", "aux"}
 	if !equalStringSlices(got, want) {
 		t.Fatalf("prefix = %#v, want %#v", got, want)
 	}
 }
 
+// TestProposedCommandPrefixRejectsPrefixLeavingUnsafeTailUncovered guards
+// against proposedCommandPrefix offering to approve one segment of a
+// multi-segment command (e.g. "ps aux") while a different segment (e.g. "npm
+// install") is not known-safe. shellExecutionArgsForApproval escalates the
+// entire command once any prefix is approved, so an uncovered unsafe segment
+// would bypass the sandbox unreviewed. Uses npm, which is never known-safe on
+// any platform, so the assertion does not depend on runtime.GOOS.
+func TestProposedCommandPrefixRejectsPrefixLeavingUnsafeTailUncovered(t *testing.T) {
+	if got := proposedCommandPrefix("bash", map[string]any{"command": "ps aux && npm install"}); got != nil {
+		t.Fatalf("expected no prefix because npm segment is not known-safe, got %#v", got)
+	}
+}
+
 func TestProposedCommandPrefixHonorsRequestedPrefixAcrossSegments(t *testing.T) {
 	got := proposedCommandPrefix("bash", map[string]any{
 		"command":     "git status --short && git status --branch",
@@ -120,13 +147,35 @@ func TestProposedCommandPrefixRejectsUnsafeLaunchers(t *testing.T) {
 func TestMatchCommandPrefixCoversSegmentedCommandWithSafeTail(t *testing.T) {
 	engine := sandbox.NewEngine(sandbox.EngineOptions{WorkspaceRoot: t.TempDir()})
 	engine.GrantCommandPrefixForSession("bash", []string{"ps", "aux"})
+	// head is MSYS-prone on Windows (#458) and must not count as a known-safe tail.
+	command := "ps aux | head -5"
+	if runtime.GOOS == "windows" {
+		command = "ps aux | echo ok"
+	}
 
-	grant, ok, session := matchCommandPrefix("bash", map[string]any{"command": "ps aux | head -5"}, Options{Sandbox: engine})
+	grant, ok, session := matchCommandPrefix("bash", map[string]any{"command": command}, Options{Sandbox: engine})
 	if !ok || !session || !equalStringSlices(grant.Prefix, []string{"ps", "aux"}) {
 		t.Fatalf("match = %#v ok=%v session=%v, want session ps aux prefix", grant, ok, session)
 	}
 }
 
+func TestKnownSafeCommandSegmentRejectsMsysProneOnWindows(t *testing.T) {
+	if runtime.GOOS != "windows" {
+		t.Skip("windows-only known-safe MSYS guard")
+	}
+	for _, command := range [][]string{{"head", "-5"}, {"cat", "file.txt"}, {"grep", "pat"}} {
+		if knownSafeCommandSegment(command) {
+			t.Fatalf("expected %q to be unsafe on Windows, got known-safe", command)
+		}
+	}
+	if !knownSafeCommandSegment([]string{"echo", "ok"}) {
+		t.Fatal("expected echo to remain known-safe on Windows")
+	}
+	if !tools.MsysProneCommandName("head") {
+		t.Fatal("expected head to be MSYS-prone")
+	}
+}
+
 func TestMatchCommandPrefixRejectsUncoveredSegment(t *testing.T) {
 	engine := sandbox.NewEngine(sandbox.EngineOptions{WorkspaceRoot: t.TempDir()})
 	engine.GrantCommandPrefixForSession("bash", []string{"ps", "aux"})

internal/agent/file_diagnostics.go

@@ -0,0 +1,62 @@
+package agent
+
+import (
+	"context"
+	"os"
+	"path/filepath"
+	"strings"
+	"time"
+
+	"github.com/Gitlawb/zero/internal/lsp"
+)
+
+// fileDiagnosticsTimeout bounds one inline post-edit diagnostics check so a
+// slow or wedged language server can never hang a tool call; on timeout the
+// edit simply reports without a diagnostics block.
+const fileDiagnosticsTimeout = 10 * time.Second
+
+// NewFileDiagnostics adapts an *lsp.Manager to the per-edit inline diagnostics
+// callback (tools.RunOptions.Diagnostics): it reads the just-written file,
+// checks it against the file's language server, and formats error-severity
+// diagnostics for the model. Warnings and hints are excluded — nagging about
+// style on every edit is noise, while a type error the edit just introduced is
+// exactly what the model should see before its next step. Diagnostics are
+// rendered with workspace-relative paths: the absolute path would put the
+// local username/home directory into the model prompt and session transcript
+// on every edit. Returns nil when manager is nil, disabling inline diagnostics
+// entirely.
+func NewFileDiagnostics(manager *lsp.Manager, workspaceRoot string) func(context.Context, string) string {
+	if manager == nil {
+		return nil
+	}
+	return func(ctx context.Context, absPath string) string {
+		text, err := os.ReadFile(absPath)
+		if err != nil {
+			return ""
+		}
+		checkCtx, cancel := context.WithTimeout(ctx, fileDiagnosticsTimeout)
+		defer cancel()
+		diagnostics, err := manager.Check(checkCtx, absPath, string(text))
+		if err != nil {
+			return ""
+		}
+		errors := lsp.FilterBySeverity(diagnostics, lsp.SeverityError)
+		if len(errors) == 0 {
+			return ""
+		}
+		return lsp.FormatDiagnostics(diagnosticsDisplayPath(workspaceRoot, absPath), errors)
+	}
+}
+
+// diagnosticsDisplayPath renders absPath relative to the workspace root for
+// model-facing output, falling back to the file's base name when the path is
+// outside the workspace (a bare name still identifies the file without
+// exposing the directory layout).
+func diagnosticsDisplayPath(workspaceRoot, absPath string) string {
+	if workspaceRoot != "" {
+		if rel, err := filepath.Rel(workspaceRoot, absPath); err == nil && !strings.HasPrefix(rel, "..") {
+			return rel
+		}
+	}
+	return filepath.Base(absPath)
+}

internal/agent/file_diagnostics_test.go

@@ -0,0 +1,24 @@
+package agent
+
+import (
+	"path/filepath"
+	"testing"
+)
+
+// Diagnostics are model-facing: absolute paths would leak the local username
+// and directory layout into the prompt and session transcript on every edit.
+func TestDiagnosticsDisplayPath(t *testing.T) {
+	root := filepath.Join("/Users", "someone", "project")
+	cases := []struct {
+		root, abs, want string
+	}{
+		{root, filepath.Join(root, "internal", "a.go"), filepath.Join("internal", "a.go")},
+		{root, filepath.Join("/etc", "other.go"), "other.go"}, // outside root -> base name only
+		{"", filepath.Join("/home", "user", "x.go"), "x.go"},  // no root -> base name only
+	}
+	for _, c := range cases {
+		if got := diagnosticsDisplayPath(c.root, c.abs); got != c.want {
+			t.Errorf("diagnosticsDisplayPath(%q, %q) = %q, want %q", c.root, c.abs, got, c.want)
+		}
+	}
+}

internal/agent/guardrails.go

@@ -429,8 +429,9 @@ func (state *guardState) observeToolResult(name string, failed bool, output stri
 func (state *guardState) observeTurn(collected zeroruntime.CollectedStream) (stop bool) {
 	hasToolCalls := len(collected.ToolCalls) > 0
 	hasVisibleText := strings.TrimSpace(collected.Text) != ""
+	hasReasoning := collected.HasReasoning || len(collected.ReasoningBlocks) > 0
 
-	if hasToolCalls || hasVisibleText {
+	if hasToolCalls || hasVisibleText || hasReasoning {
 		state.emptyTurns = 0
 	} else {
 		state.emptyTurns++

internal/agent/guardrails_test.go

@@ -22,6 +22,14 @@ func textTurn(content string) []zeroruntime.StreamEvent {
 	}
 }
 
+// reasoningTurn produces live reasoning without visible assistant text.
+func reasoningTurn(content string) []zeroruntime.StreamEvent {
+	return []zeroruntime.StreamEvent{
+		{Type: zeroruntime.StreamEventReasoning, Content: content},
+		{Type: zeroruntime.StreamEventDone},
+	}
+}
+
 // toolTurn produces a turn that calls a named tool with the given args JSON.
 func toolTurn(callID string, toolName string, args string) []zeroruntime.StreamEvent {
 	return []zeroruntime.StreamEvent{
@@ -101,6 +109,31 @@ func TestRunResetsEmptyTurnCounterOnVisibleOutput(t *testing.T) {
 	}
 }
 
+func TestRunResetsEmptyTurnCounterOnReasoning(t *testing.T) {
+	provider := &mockProvider{
+		turns: [][]zeroruntime.StreamEvent{
+			reasoningTurn("thinking 1"),
+			reasoningTurn("thinking 2"),
+			reasoningTurn("thinking 3"),
+			textTurn("done"),
+		},
+	}
+
+	result, err := Run(context.Background(), "go", provider, Options{
+		Registry: tools.NewRegistry(),
+		MaxTurns: 12,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if result.FinalAnswer != "done" {
+		t.Fatalf("expected reasoning-only turns to keep the run live until final answer, got %q", result.FinalAnswer)
+	}
+	if len(provider.requests) != 4 {
+		t.Fatalf("expected 4 turns, got %d", len(provider.requests))
+	}
+}
+
 func TestRunResetsEmptyTurnCounterOnToolCall(t *testing.T) {
 	root := t.TempDir()
 	writeAgentTestFile(t, root+"/notes.txt", "alpha")

internal/agent/loop.go

@@ -137,6 +137,12 @@ func Run(ctx context.Context, prompt string, provider Provider, options Options)
 	guards := newGuardState()
 	compactor := newCompactionState(options)
 
+	// Background post-edit diagnostics: files changed by mutating tools are
+	// checked off the tool-call critical path and any errors are appended as a
+	// nudge before the NEXT request (see async_diagnostics.go). nil when no
+	// FileDiagnostics callback is wired; every method no-ops on nil.
+	postEditDiagnostics := newAsyncDiagnostics(options.FileDiagnostics, options.Cwd)
+
 	// loaded tracks deferred-eligible tools the model has pulled via tool_search
 	// during THIS run. It is consulted by partitionTools each turn to expose a
 	// loaded tool's full schema; it lives only for the run (v1 within-run scope).
@@ -159,6 +165,17 @@ func Run(ctx context.Context, prompt string, provider Provider, options Options)
 	for turn := 0; turn < maxTurns; turn++ {
 		result.Turns = turn + 1
 
+		// Deliver background post-edit diagnostics from the previous turn's edits
+		// BEFORE compaction so the nudge is part of the request being budgeted.
+		// A brief wait at most (asyncDiagnosticsDrainTimeout); an unfinished check
+		// simply delivers on a later turn.
+		if nudge := postEditDiagnostics.drain(ctx); nudge != "" {
+			messages = append(messages, zeroruntime.Message{
+				Role:    zeroruntime.MessageRoleUser,
+				Content: nudge,
+			})
+		}
+
 		// Build the per-turn tool list first so proactive compaction can include
 		// the tool-definition tokens (they ride on every request) in its estimate.
 		// partitionTools depends only on registry/permissions/options/loaded, not on
@@ -173,6 +190,7 @@ func Run(ctx context.Context, prompt string, provider Provider, options Options)
 			Messages:        copyMessages(messages),
 			Tools:           exposed,
 			ReasoningEffort: options.ReasoningEffort,
+			PromptCacheKey:  options.SessionID,
 		}
 
 		// Report the per-category context budget for this turn so a surface can
@@ -208,6 +226,7 @@ func Run(ctx context.Context, prompt string, provider Provider, options Options)
 					Messages:        copyMessages(messages),
 					Tools:           exposed,
 					ReasoningEffort: options.ReasoningEffort,
+					PromptCacheKey:  options.SessionID,
 				}
 				// Pre-content connect after a context-limit compaction: route through the
 				// reconnect helper so a transient upstream hiccup here doesn't fail the
@@ -268,6 +287,7 @@ func Run(ctx context.Context, prompt string, provider Provider, options Options)
 					Messages:        copyMessages(messages),
 					Tools:           exposed,
 					ReasoningEffort: options.ReasoningEffort,
+					PromptCacheKey:  options.SessionID,
 				}
 				retryStream, retryStreamErr := streamWithReconnect(ctx, provider, retryRequest, reconnectNoticeFor(options))
 				if retryStreamErr != nil {
@@ -326,6 +346,7 @@ func Run(ctx context.Context, prompt string, provider Provider, options Options)
 				Messages:        copyMessages(messages),
 				Tools:           exposed,
 				ReasoningEffort: options.ReasoningEffort,
+				PromptCacheKey:  options.SessionID,
 			}
 			retryStream, retryErr := streamWithReconnect(ctx, provider, retryRequest, reconnectNoticeFor(options))
 			if retryErr != nil {
@@ -470,6 +491,19 @@ func Run(ctx context.Context, prompt string, provider Provider, options Options)
 					continue
 				}
 			}
+			// Finalization diagnostics gate: edits from this run may still have
+			// checks in flight — the per-turn drain waits only briefly and defers
+			// to "a later turn", but a final answer means there is no later turn.
+			// Wait out the full inline-era budget once; an introduced error gives
+			// the model one more turn to see (and fix) it instead of being lost.
+			// Free for runs that never edited: an idle collector returns "".
+			if nudge := postEditDiagnostics.drainFinal(ctx); nudge != "" {
+				messages = append(messages, zeroruntime.Message{
+					Role:    zeroruntime.MessageRoleUser,
+					Content: nudge,
+				})
+				continue
+			}
 			result.FinalAnswer = collected.Text
 			result.Messages = copyMessages(messages)
 			return result, nil
@@ -495,11 +529,35 @@ func Run(ctx context.Context, prompt string, provider Provider, options Options)
 		// between tool_results breaks strict provider replay) — same after-batch
 		// rationale as turnRequestedModel above.
 		var changedFilesThisBatch []string
+		// Parallel read-ahead state: results for calls[precomputedStart:precomputedEnd]
+		// executed concurrently, consumed strictly in order below.
+		var precomputed []precomputedToolResult
+		precomputedStart, precomputedEnd := 0, 0
 		for index, call := range collected.ToolCalls {
+			// When this call starts a consecutive run of >= 2 auto-allowed read-only
+			// calls, execute the whole run concurrently now (see parallel_tools.go).
+			// The scan happens lazily at the run's first index — never ahead of a
+			// pending mutating call — so read-after-write ordering is preserved.
+			if index >= precomputedEnd {
+				runEnd := index
+				for runEnd < len(collected.ToolCalls) && parallelSafeToolCall(registry, collected.ToolCalls[runEnd], options) {
+					runEnd++
+				}
+				if runEnd-index >= 2 {
+					precomputed = executeParallelReadBatch(ctx, registry, collected.ToolCalls, index, runEnd, permissionMode, options)
+					precomputedStart, precomputedEnd = index, runEnd
+				}
+			}
 			if options.OnToolCall != nil {
 				options.OnToolCall(call)
 			}
-			toolResult, abortErr := executeToolCall(ctx, registry, call, permissionMode, options)
+			var toolResult ToolResult
+			var abortErr error
+			if index >= precomputedStart && index < precomputedEnd {
+				toolResult, abortErr = precomputed[index-precomputedStart].result, precomputed[index-precomputedStart].abortErr
+			} else {
+				toolResult, abortErr = executeToolCall(ctx, registry, call, permissionMode, options)
+			}
 			if options.OnToolResult != nil {
 				options.OnToolResult(toolResult)
 			}
@@ -562,11 +620,14 @@ func Run(ctx context.Context, prompt string, provider Provider, options Options)
 				failureHint = toolFailureHint(call.Name, toolSchemaJSON(registry, call.Name), toolResult.Output)
 			}
 
-			// Post-edit self-correction: collect the files this successful mutating
-			// tool changed; verification runs once over the union after the batch.
-			// A read-only tool (no ChangedFiles) never contributes.
-			if options.SelfCorrect != nil && toolResult.Status == tools.StatusOK && len(toolResult.ChangedFiles) > 0 {
+			// Collect the files this successful mutating tool changed. Self-correct
+			// verification runs once over the union after the batch; background
+			// diagnostics start NOW so the language server analyzes while the rest
+			// of the batch executes. A read-only tool (no ChangedFiles) never
+			// contributes.
+			if toolResult.Status == tools.StatusOK && len(toolResult.ChangedFiles) > 0 {
 				changedFilesThisBatch = append(changedFilesThisBatch, toolResult.ChangedFiles...)
+				postEditDiagnostics.enqueue(ctx, toolResult.ChangedFiles)
 			}
 		}
 
@@ -649,6 +710,17 @@ func Run(ctx context.Context, prompt string, provider Provider, options Options)
 	// mid-run, not a natural completion (a finished run returns via the no-tool-call
 	// success path before this). Under the headless completion gate that is INCOMPLETE,
 	// not success, so a run that loops to the turn limit isn't reported as done.
+	//
+	// The final-answer call is one more model request, so diagnostics from the
+	// LAST turn's edits get a drain too — with the finalization budget, since
+	// there is no later turn to defer to — otherwise an error introduced by
+	// the final edit would go unreported in the summary.
+	if nudge := postEditDiagnostics.drainFinal(ctx); nudge != "" {
+		messages = append(messages, zeroruntime.Message{
+			Role:    zeroruntime.MessageRoleUser,
+			Content: nudge,
+		})
+	}
 	if answer, finalMessages, finishReason := finalAnswerAfterMaxTurns(ctx, provider, messages, options); strings.TrimSpace(answer) != "" {
 		result.FinalAnswer = answer
 		result.FinishReason = finishReason
@@ -681,6 +753,7 @@ func finalAnswerAfterMaxTurns(ctx context.Context, provider Provider, messages [
 	stream, err := streamWithReconnect(ctx, provider, zeroruntime.CompletionRequest{
 		Messages:        copyMessages(finalMessages),
 		ReasoningEffort: options.ReasoningEffort,
+		PromptCacheKey:  options.SessionID,
 	}, reconnectNoticeFor(options))
 	if err != nil {
 		return "", messages, ""
@@ -1071,6 +1144,12 @@ func executeToolCall(ctx context.Context, registry *tools.Registry, call ToolCal
 		// Per-session file version tracker so write_file/edit_file refuse to clobber
 		// a file that changed on disk outside Zero since it was last read.
 		FileTracker: options.FileTracker,
+		// Post-edit diagnostics are NOT forwarded to the tools: they used to run
+		// synchronously inside edit_file/write_file and block every mutating tool
+		// call on the language server (≥300ms debounce floor, 10s cap). The loop
+		// now checks changed files in the background and nudges the model before
+		// its next request instead (see async_diagnostics.go). The tools' inline
+		// mechanism (tools.RunOptions.Diagnostics) stays for direct API callers.
 		// Forward the run's operator tool filters so a filter-aware tool
 		// (tool_search) never discloses or loads an operator-hidden deferred tool.
 		EnabledTools:  options.EnabledTools,

internal/agent/loop_test.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"encoding/json"
 	"errors"
+	"fmt"
 	"os"
 	"path/filepath"
 	"runtime"
@@ -1945,6 +1946,11 @@ func TestRunCommandPrefixApprovalBypassesSandboxForMatchingShellCalls(t *testing
 
 func TestRunCommandPrefixApprovalCoversSegmentedShellWithSafeTail(t *testing.T) {
 	root := t.TempDir()
+	segmentedCommand := `ps aux | head -5`
+	if runtime.GOOS == "windows" {
+		// head is MSYS-prone on Windows (#458) and no longer counts as a known-safe tail.
+		segmentedCommand = `ps aux | echo ok`
+	}
 	retryTool := &sandboxDeniedRetryTool{}
 	registry := tools.NewRegistry()
 	registry.Register(retryTool)
@@ -1958,7 +1964,7 @@ func TestRunCommandPrefixApprovalCoversSegmentedShellWithSafeTail(t *testing.T)
 			},
 			{
 				{Type: zeroruntime.StreamEventToolCallStart, ToolCallID: "call-2", ToolName: "bash"},
-				{Type: zeroruntime.StreamEventToolCallDelta, ToolCallID: "call-2", ArgumentsFragment: `{"command":"ps aux | head -5"}`},
+				{Type: zeroruntime.StreamEventToolCallDelta, ToolCallID: "call-2", ArgumentsFragment: fmt.Sprintf(`{"command":%q}`, segmentedCommand)},
 				{Type: zeroruntime.StreamEventToolCallEnd, ToolCallID: "call-2"},
 				{Type: zeroruntime.StreamEventDone},
 			},
@@ -2907,8 +2913,10 @@ func TestBuildSystemPromptInjectsHostShellContext(t *testing.T) {
 		t.Fatalf("expected operating system in environment block, got %q", prompt)
 	}
 	if runtime.GOOS == "windows" {
-		if !strings.Contains(prompt, "Windows cmd.exe syntax") || !strings.Contains(prompt, "cwd argument") {
-			t.Fatalf("expected Windows cmd.exe shell guidance in prompt, got %q", prompt)
+		for _, want := range []string{"Windows cmd.exe syntax", "cwd argument", "MSYS binaries", "grep", "require_escalated"} {
+			if !strings.Contains(prompt, want) {
+				t.Fatalf("expected Windows shell guidance to mention %q, got %q", want, prompt)
+			}
 		}
 	} else if !strings.Contains(prompt, "/bin/sh syntax") {
 		t.Fatalf("expected POSIX shell guidance in prompt, got %q", prompt)

internal/agent/parallel_tools.go

@@ -0,0 +1,97 @@
+package agent
+
+import (
+	"context"
+	"sync"
+
+	"github.com/Gitlawb/zero/internal/tools"
+)
+
+// Parallel read-ahead for tool batches. When a turn requests several
+// independent lookups (read_file + grep + glob is the common shape), executing
+// them one after another serializes pure I/O waits. A consecutive run of
+// auto-allowed read-only calls is executed concurrently instead; results are
+// then consumed in the original call order, so guard counters, message
+// ordering, abort semantics, and the surface's call/result event pairing are
+// byte-identical to sequential execution. Runs never span a mutating call: a
+// read that follows a write must observe the write, so eligibility is decided
+// per consecutive run, not per batch.
+
+// maxParallelReadTools bounds concurrent read-only tool executions in a turn.
+const maxParallelReadTools = 8
+
+// precomputedToolResult is one parallel read-ahead execution, keyed back to
+// its batch index by the caller.
+type precomputedToolResult struct {
+	result   ToolResult
+	abortErr error
+}
+
+// parallelSafeToolCall reports whether call may run concurrently with its
+// neighbors: the tool must exist, be side-effect-free (SideEffectRead), and be
+// auto-allowed for these args, so no interactive prompt or workspace mutation
+// is on the hot path. Loop-intercepted tools (ask_user, request_permissions)
+// and tool_search (mutates the deferred-tool set) stay sequential.
+func parallelSafeToolCall(registry *tools.Registry, call ToolCall, options Options) bool {
+	switch call.Name {
+	case "ask_user", tools.RequestPermissionsToolName, tools.ToolSearchToolName:
+		return false
+	}
+	tool, found := registry.Get(call.Name)
+	if !found || tool.Safety().SideEffect != tools.SideEffectRead {
+		return false
+	}
+	args := map[string]any{}
+	if call.Arguments != "" {
+		if err := decodeToolArguments(call.Arguments, &args); err != nil {
+			return false
+		}
+	}
+	return effectivePermission(tool, args) == tools.PermissionAllow
+}
+
+// executeParallelReadBatch runs calls[start:end] concurrently (bounded by
+// maxParallelReadTools) and returns results indexed relative to start. All
+// execution-side callbacks that can fire inside executeToolCall are serialized
+// behind one mutex: a permission prompt (a sandbox preflight can demand one
+// even for an auto-allowed read) must never appear twice at once on an
+// interactive front-end, and OnPermission event handlers append to shared
+// session-recording state without their own locking — two batched reads under
+// a granted extra root would otherwise race (the pre-batch serial loop never
+// had two callbacks in flight at once).
+func executeParallelReadBatch(ctx context.Context, registry *tools.Registry, calls []ToolCall, start, end int, permissionMode PermissionMode, options Options) []precomputedToolResult {
+	batchOptions := options
+	var callbackMutex sync.Mutex
+	if options.OnPermissionRequest != nil {
+		inner := options.OnPermissionRequest
+		batchOptions.OnPermissionRequest = func(ctx context.Context, request PermissionRequest) (PermissionDecision, error) {
+			callbackMutex.Lock()
+			defer callbackMutex.Unlock()
+			return inner(ctx, request)
+		}
+	}
+	if options.OnPermission != nil {
+		inner := options.OnPermission
+		batchOptions.OnPermission = func(event PermissionEvent) {
+			callbackMutex.Lock()
+			defer callbackMutex.Unlock()
+			inner(event)
+		}
+	}
+
+	results := make([]precomputedToolResult, end-start)
+	semaphore := make(chan struct{}, maxParallelReadTools)
+	var waitGroup sync.WaitGroup
+	for index := start; index < end; index++ {
+		waitGroup.Add(1)
+		go func(index int) {
+			defer waitGroup.Done()
+			semaphore <- struct{}{}
+			defer func() { <-semaphore }()
+			result, abortErr := executeToolCall(ctx, registry, calls[index], permissionMode, batchOptions)
+			results[index-start] = precomputedToolResult{result: result, abortErr: abortErr}
+		}(index)
+	}
+	waitGroup.Wait()
+	return results
+}

internal/agent/parallel_tools_test.go

@@ -0,0 +1,204 @@
+package agent
+
+import (
+	"context"
+	"fmt"
+	"sync"
+	"testing"
+	"time"
+
+	"github.com/Gitlawb/zero/internal/tools"
+	"github.com/Gitlawb/zero/internal/zeroruntime"
+)
+
+// probeTool records execution overlap and ordering so tests can assert what
+// actually ran concurrently.
+type probeTool struct {
+	name       string
+	sideEffect tools.SideEffect
+	delay      time.Duration
+	// shared, when set, receives this probe's start/end entries too, so a test
+	// can observe ordering ACROSS probes (e.g. reads vs a write barrier).
+	shared *probeLog
+
+	mu        sync.Mutex
+	active    int
+	maxActive int
+	log       []string
+}
+
+// probeLog is a mutex-guarded event log shared between probes.
+type probeLog struct {
+	mu      sync.Mutex
+	entries []string
+}
+
+func (log *probeLog) append(entry string) {
+	log.mu.Lock()
+	defer log.mu.Unlock()
+	log.entries = append(log.entries, entry)
+}
+
+func (log *probeLog) snapshot() []string {
+	log.mu.Lock()
+	defer log.mu.Unlock()
+	return append([]string(nil), log.entries...)
+}
+
+func (tool *probeTool) Name() string        { return tool.name }
+func (tool *probeTool) Description() string { return "test probe tool" }
+func (tool *probeTool) Parameters() tools.Schema {
+	return tools.Schema{
+		Type:                 "object",
+		Properties:           map[string]tools.PropertySchema{"id": {Type: "string"}},
+		AdditionalProperties: false,
+	}
+}
+func (tool *probeTool) Safety() tools.Safety {
+	return tools.Safety{SideEffect: tool.sideEffect, Permission: tools.PermissionAllow, Reason: "test"}
+}
+func (tool *probeTool) Run(_ context.Context, args map[string]any) tools.Result {
+	id, _ := args["id"].(string)
+	tool.mu.Lock()
+	tool.active++
+	if tool.active > tool.maxActive {
+		tool.maxActive = tool.active
+	}
+	tool.log = append(tool.log, "start:"+id)
+	tool.mu.Unlock()
+	if tool.shared != nil {
+		tool.shared.append("start:" + id)
+	}
+	time.Sleep(tool.delay)
+	if tool.shared != nil {
+		tool.shared.append("end:" + id)
+	}
+	tool.mu.Lock()
+	tool.active--
+	tool.log = append(tool.log, "end:"+id)
+	tool.mu.Unlock()
+	return tools.Result{Status: tools.StatusOK, Output: "probe " + id}
+}
+
+func probeCallEvents(callID, toolName, id string) []zeroruntime.StreamEvent {
+	return []zeroruntime.StreamEvent{
+		{Type: zeroruntime.StreamEventToolCallStart, ToolCallID: callID, ToolName: toolName},
+		{Type: zeroruntime.StreamEventToolCallDelta, ToolCallID: callID, ArgumentsFragment: fmt.Sprintf(`{"id":%q}`, id)},
+		{Type: zeroruntime.StreamEventToolCallEnd, ToolCallID: callID},
+	}
+}
+
+func TestParallelSafeToolCall(t *testing.T) {
+	registry := tools.NewRegistry()
+	registry.Register(&probeTool{name: "probe_read", sideEffect: tools.SideEffectRead})
+	registry.Register(&probeTool{name: "probe_write", sideEffect: tools.SideEffectWrite})
+
+	call := func(name, args string) ToolCall { return ToolCall{ID: "c", Name: name, Arguments: args} }
+	if !parallelSafeToolCall(registry, call("probe_read", `{"id":"a"}`), Options{}) {
+		t.Fatal("auto-allowed read tool must be parallel-safe")
+	}
+	if parallelSafeToolCall(registry, call("probe_write", `{"id":"a"}`), Options{}) {
+		t.Fatal("mutating tool must not be parallel-safe")
+	}
+	if parallelSafeToolCall(registry, call("unknown_tool", `{}`), Options{}) {
+		t.Fatal("unknown tool must not be parallel-safe")
+	}
+	if parallelSafeToolCall(registry, call("probe_read", `{"id":`), Options{}) {
+		t.Fatal("undecodable arguments must not be parallel-safe")
+	}
+	if parallelSafeToolCall(registry, call("ask_user", `{}`), Options{}) {
+		t.Fatal("loop-intercepted tools must stay sequential")
+	}
+}
+
+func TestRunExecutesConsecutiveReadsConcurrently(t *testing.T) {
+	probe := &probeTool{name: "probe_read", sideEffect: tools.SideEffectRead, delay: 60 * time.Millisecond}
+	registry := tools.NewRegistry()
+	registry.Register(probe)
+
+	turnOne := append(probeCallEvents("call-1", "probe_read", "a"), probeCallEvents("call-2", "probe_read", "b")...)
+	turnOne = append(turnOne, probeCallEvents("call-3", "probe_read", "c")...)
+	turnOne = append(turnOne, zeroruntime.StreamEvent{Type: zeroruntime.StreamEventDone})
+	provider := &mockProvider{
+		turns: [][]zeroruntime.StreamEvent{
+			turnOne,
+			{{Type: zeroruntime.StreamEventText, Content: "done"}, {Type: zeroruntime.StreamEventDone}},
+		},
+	}
+
+	var results []ToolResult
+	_, err := Run(context.Background(), "probe", provider, Options{
+		Registry:     registry,
+		OnToolResult: func(result ToolResult) { results = append(results, result) },
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if probe.maxActive < 2 {
+		t.Fatalf("consecutive read-only calls must overlap, max concurrency was %d", probe.maxActive)
+	}
+	// Results must still be recorded in original call order.
+	if len(results) != 3 || results[0].ToolCallID != "call-1" || results[1].ToolCallID != "call-2" || results[2].ToolCallID != "call-3" {
+		t.Fatalf("tool results out of order: %#v", results)
+	}
+	messages := provider.requests[1].Messages
+	var toolOrder []string
+	for _, message := range messages {
+		if message.Role == zeroruntime.MessageRoleTool {
+			toolOrder = append(toolOrder, message.ToolCallID)
+		}
+	}
+	if len(toolOrder) != 3 || toolOrder[0] != "call-1" || toolOrder[1] != "call-2" || toolOrder[2] != "call-3" {
+		t.Fatalf("recorded tool messages out of order: %v", toolOrder)
+	}
+}
+
+func TestRunParallelReadsNeverSpanMutatingCall(t *testing.T) {
+	shared := &probeLog{}
+	read := &probeTool{name: "probe_read", sideEffect: tools.SideEffectRead, delay: 30 * time.Millisecond, shared: shared}
+	write := &probeTool{name: "probe_write", sideEffect: tools.SideEffectWrite, shared: shared}
+	registry := tools.NewRegistry()
+	registry.Register(read)
+	registry.Register(write)
+
+	turnOne := append(probeCallEvents("call-1", "probe_read", "r1"), probeCallEvents("call-2", "probe_read", "r2")...)
+	turnOne = append(turnOne, probeCallEvents("call-3", "probe_write", "w")...)
+	turnOne = append(turnOne, probeCallEvents("call-4", "probe_read", "r3")...)
+	turnOne = append(turnOne, probeCallEvents("call-5", "probe_read", "r4")...)
+	turnOne = append(turnOne, zeroruntime.StreamEvent{Type: zeroruntime.StreamEventDone})
+	provider := &mockProvider{
+		turns: [][]zeroruntime.StreamEvent{
+			turnOne,
+			{{Type: zeroruntime.StreamEventText, Content: "done"}, {Type: zeroruntime.StreamEventDone}},
+		},
+	}
+
+	_, err := Run(context.Background(), "probe", provider, Options{Registry: registry})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	// Cross-probe ordering on the SHARED log: the write must start only after
+	// both first-batch reads finished, and both second-batch reads must start
+	// only after the write finished — batches never cross a mutating call.
+	log := shared.snapshot()
+	index := func(entry string) int {
+		for i, e := range log {
+			if e == entry {
+				return i
+			}
+		}
+		t.Fatalf("entry %q missing from shared log: %v", entry, log)
+		return -1
+	}
+	writeStart, writeEnd := index("start:w"), index("end:w")
+	if firstBatchMaxEnd := max(index("end:r1"), index("end:r2")); writeStart < firstBatchMaxEnd {
+		t.Fatalf("write started before the first read batch finished: %v", log)
+	}
+	if secondBatchMinStart := min(index("start:r3"), index("start:r4")); secondBatchMinStart < writeEnd {
+		t.Fatalf("second read batch started before the write finished: %v", log)
+	}
+	if read.maxActive < 2 {
+		t.Fatalf("reads within a batch must overlap, max concurrency was %d", read.maxActive)
+	}
+}

internal/agent/partition_cache_stable_test.go

@@ -34,7 +34,7 @@ func TestPartitionToolsActiveAppendsLoadedToolAfterEagerBlock(t *testing.T) {
 		t.Fatalf("expected read_file, tool_search and loaded alpha all exposed, got %#v", exposed)
 	}
 	// The eager tools must precede the loaded deferred tool (appended, not inserted).
-	if !(readPos < alphaPos && searchPos < alphaPos) {
+	if readPos >= alphaPos || searchPos >= alphaPos {
 		t.Fatalf("loaded deferred tool must be appended after the eager block; got read=%d search=%d alpha=%d", readPos, searchPos, alphaPos)
 	}
 	// beta was never loaded — it stays hidden.

internal/agent/skills_context_test.go

@@ -7,14 +7,18 @@ import (
 )
 
 func TestSkillsContextCapsLongList(t *testing.T) {
-	skills := make([]SkillInfo, 0, 40)
-	for i := 0; i < 40; i++ {
+	// 60 skills with realistic trigger-rich descriptions (~180 chars each, under
+	// the 200-rune truncation) total well past the 4096-byte list budget, so the
+	// overflow summary must kick in while the block stays bounded.
+	longDesc := strings.Repeat("use when the request touches deployments, release notes, or verification; ", 2) + "number "
+	skills := make([]SkillInfo, 0, 60)
+	for i := 0; i < 60; i++ {
 		n := strconv.Itoa(i)
-		skills = append(skills, SkillInfo{Name: "skill-" + n, Description: "does something useful, number " + n})
+		skills = append(skills, SkillInfo{Name: "skill-" + n, Description: longDesc + n})
 	}
 	got := skillsContext(Options{Skills: skills})
-	if len(got) > 1200 {
-		t.Fatalf("skills block should stay bounded, got %d bytes:\n%s", len(got), got)
+	if len(got) > skillsContextListBudget+1000 {
+		t.Fatalf("skills block should stay bounded near the %d-byte budget, got %d bytes:\n%s", skillsContextListBudget, len(got), got)
 	}
 	if !strings.Contains(got, "more (call skill") {
 		t.Fatalf("expected an overflow summary line, got:\n%s", got)
@@ -25,6 +29,27 @@ func TestSkillsContextCapsLongList(t *testing.T) {
 	}
 }
 
+// A realistic mid-size skill set (20 skills, trigger-rich descriptions) must be
+// listed in FULL — no overflow summary. This pins the fix for the old 640-byte
+// budget, under which skills past ~#6 were invisible to the model and therefore
+// never triggered.
+func TestSkillsContextListsRealisticSetInFull(t *testing.T) {
+	desc := "Use when the user asks about deployments, release notes, or pre-merge verification runs."
+	skills := make([]SkillInfo, 0, 20)
+	for i := 0; i < 20; i++ {
+		skills = append(skills, SkillInfo{Name: "skill-" + strconv.Itoa(i), Description: desc})
+	}
+	got := skillsContext(Options{Skills: skills})
+	if strings.Contains(got, "more (call skill") {
+		t.Fatalf("20 described skills must all be listed without overflow, got:\n%s", got)
+	}
+	for i := 0; i < 20; i++ {
+		if !strings.Contains(got, "- skill-"+strconv.Itoa(i)+":") {
+			t.Fatalf("skill-%d missing from the list:\n%s", i, got)
+		}
+	}
+}
+
 func TestSkillsContext(t *testing.T) {
 	if got := skillsContext(Options{}); got != "" {
 		t.Fatalf("no skills should yield an empty section, got %q", got)

internal/agent/system_prompt.go

@@ -9,6 +9,7 @@ import (
 	"strings"
 	"unicode/utf8"
 
+	"github.com/Gitlawb/zero/internal/config"
 	"github.com/Gitlawb/zero/internal/repomap"
 	"github.com/Gitlawb/zero/internal/workspaceseed"
 )
@@ -38,6 +39,14 @@ const fallbackSystemPrompt = "You are Zero, a terminal coding agent. Help with t
 // general-to-specific order.
 var projectContextFiles = []string{"AGENTS.md", "ZERO.md", ".zero/AGENTS.md"}
 
+// userContextFile is the per-user instruction file, resolved under
+// config.UserConfigDir()/zero/ alongside the rest of Zero's per-user config
+// (config.json, commands, specialists) so users can keep personal guidance out
+// of individual repositories.
+const userContextFile = "ZERO.md"
+
+var userConfigDirForPrompt = config.UserConfigDir
+
 // maxProjectContextBytes caps how much of a single project doc is injected so
 // a large guidelines file can't blow the context budget.
 const maxProjectContextBytes = 8 << 10 // 8 KiB per file
@@ -81,6 +90,13 @@ func buildSystemPrompt(options Options) string {
 	if seed := workspaceSeedContext(options.Cwd); seed != "" {
 		sections = append(sections, seed)
 	}
+	// User guidelines are injected before workspace/project guidelines so the
+	// project's AGENTS.md/ZERO.md is the later, more specific instruction
+	// block. See userGuidelines for the explicit precedence note carried in
+	// the section text itself.
+	if user := userGuidelines(); user != "" {
+		sections = append(sections, user)
+	}
 	if ws := workspaceContext(options.Cwd); ws != "" {
 		sections = append(sections, ws)
 	}
@@ -175,10 +191,19 @@ func specialistDelegationContext(options Options) string {
 // when no skills are installed, so a skill-less run reproduces the previous prompt
 // byte-for-byte.
 // skillsContextListBudget bounds the bytes spent listing individual skills so a
-// workspace with dozens of them can't bloat every turn's prompt. Skills past the
-// budget are summarized as a count rather than dropped — the model can still load
-// any of them by name (an unknown name returns the full list from the skill tool).
-const skillsContextListBudget = 640
+// workspace with an extreme number of them can't bloat every turn's prompt.
+// Skills past the budget are summarized as a count rather than dropped — the
+// model can still load any of them by name (an unknown name returns the full
+// list from the skill tool).
+//
+// The budget is deliberately generous — roughly 18 skills with maximally long
+// (200-rune, ASCII) descriptions, or 40+ with typical terser ones; non-ASCII
+// descriptions consume it faster since the budget counts bytes. This list is
+// the model's ONLY discovery surface, so a skill squeezed out of it effectively
+// never triggers. The previous 640-byte cap silently broke invocation for
+// anyone with more than ~6 skills — the model cannot match a request against a
+// skill it cannot see.
+const skillsContextListBudget = 4096
 
 func skillsContext(options Options) string {
 	if len(options.Skills) == 0 {
@@ -186,7 +211,7 @@ func skillsContext(options Options) string {
 	}
 	var b strings.Builder
 	b.WriteString("<available_skills>\n")
-	b.WriteString("Reusable, on-demand instruction sets you can load with the skill tool. When a request matches one, call skill with its exact name to pull in its guidance before acting — do not guess names or skip a relevant skill.\n")
+	b.WriteString("Reusable, on-demand instruction sets you can load with the skill tool. Before acting on a request, scan this list; when the request matches a skill's name or description, call skill with its exact name FIRST and follow the loaded guidance — do not guess names, do not skip a matching skill, and do not substitute your own approach for its instructions.\n")
 	listed, spent, omitted := 0, 0, 0
 	for _, info := range options.Skills {
 		name := strings.TrimSpace(info.Name)
@@ -216,9 +241,11 @@ func skillsContext(options Options) string {
 }
 
 // truncateForSkillLine keeps a skill's one-line description short so a single
-// verbose description can't dominate the skills-list budget.
+// verbose description can't dominate the skills-list budget. The cap is roomy on
+// purpose: the description carries the skill's trigger conditions ("use when…"),
+// and truncating those is what stops the model from ever invoking the skill.
 func truncateForSkillLine(desc string) string {
-	const maxDescRunes = 100
+	const maxDescRunes = 200
 	runes := []rune(desc)
 	if len(runes) <= maxDescRunes {
 		return desc
@@ -259,7 +286,7 @@ func workspaceContext(cwd string) string {
 	b.WriteString("Working directory: " + cwd + "\n")
 	b.WriteString("Operating system: " + runtime.GOOS + "\n")
 	if runtime.GOOS == "windows" {
-		b.WriteString("Shell syntax: Windows cmd.exe syntax for exec_command/bash tools; prefer the workdir/cwd argument instead of cd when changing directories.\n")
+		b.WriteString("Shell syntax: Windows cmd.exe syntax for exec_command/bash tools; prefer the workdir/cwd argument instead of cd when changing directories. Do not pipe to or invoke POSIX coreutils from Git for Windows (usr\\bin head/grep/tail/cat/...): they are MSYS binaries and fail under the write-restricted sandbox; use native Zero tools (grep, read_file, list_directory, glob) or cmd.exe findstr/more instead, or sandbox_permissions require_escalated only when host-level execution is truly required.\n")
 	} else {
 		b.WriteString("Shell syntax: /bin/sh syntax for exec_command/bash tools; prefer the workdir/cwd argument instead of cd when changing directories.\n")
 	}
@@ -312,13 +339,7 @@ func projectGuidelines(cwd, gitRoot string) string {
 		if remaining := maxProjectContextTotalBytes - totalUsed; remaining < limit {
 			limit = remaining
 		}
-		if len(content) > limit {
-			cut := limit
-			for cut > 0 && !utf8.RuneStart(content[cut]) {
-				cut--
-			}
-			content = content[:cut] + "\n… (truncated)"
-		}
+		content = truncateGuidelineContent(content, limit)
 		label := projectGuidelineLabel(match, gitRoot)
 		b.WriteString("\n\n## Project guidelines (" + label + ")\n\n" + content)
 		totalUsed += len(content)
@@ -326,6 +347,92 @@ func projectGuidelines(cwd, gitRoot string) string {
 	return b.String()
 }
 
+// userGuidelines returns the per-user ZERO.md instructions block, if present.
+// The file lives in config.UserConfigDir()/zero/ next to Zero's other
+// per-user config; the basename match is case-insensitive so a file saved as
+// zero.md still resolves on case-sensitive filesystems, mirroring the project
+// guideline loader. The section carries an explicit precedence note because
+// this is a global, personal preferences file: it is injected earlier in the
+// prompt than the project's AGENTS.md/ZERO.md (see buildSystemPrompt), and
+// the note keeps that precedence unambiguous even if a model weighs later
+// context more heavily than section order alone implies.
+func userGuidelines() string {
+	configDir, err := userConfigDirForPrompt()
+	if err != nil {
+		return ""
+	}
+	configDir = strings.TrimSpace(configDir)
+	if configDir == "" {
+		return ""
+	}
+	match := findCaseInsensitiveFile(filepath.Join(configDir, "zero"), userContextFile)
+	if match == "" {
+		return ""
+	}
+	data, err := os.ReadFile(match)
+	if err != nil {
+		return ""
+	}
+	content := strings.TrimSpace(string(data))
+	if content == "" {
+		return ""
+	}
+	content = truncateGuidelineContent(content, maxProjectContextBytes)
+	return "## User guidelines (" + filepath.Base(match) + ")\n\n" +
+		"These are the operator's personal preferences, not project policy. " +
+		"Where they conflict with a repository's project guidelines below (AGENTS.md/ZERO.md), the project guidelines take precedence.\n\n" +
+		content
+}
+
+// truncationMarker is appended to guideline content that was cut short.
+const truncationMarker = "\n… (truncated)"
+
+// truncateGuidelineContent caps content at limit bytes without splitting a
+// UTF-8 rune, appending a truncation marker when anything was cut. Space for
+// the marker is reserved before choosing the cut point. When limit is too
+// small to fit the marker itself, the marker is dropped instead, so the
+// returned string never exceeds limit for any non-negative limit.
+func truncateGuidelineContent(content string, limit int) string {
+	if len(content) <= limit {
+		return content
+	}
+	if limit <= 0 {
+		return ""
+	}
+	cut := limit - len(truncationMarker)
+	if cut < 0 {
+		cut = 0
+	}
+	for cut > 0 && !utf8.RuneStart(content[cut]) {
+		cut--
+	}
+	if truncated := content[:cut] + truncationMarker; len(truncated) <= limit {
+		return truncated
+	}
+	// limit is smaller than the marker itself: fall back to a hard,
+	// rune-safe cut at limit bytes with no marker rather than exceeding it.
+	end := limit
+	for end > 0 && !utf8.RuneStart(content[end]) {
+		end--
+	}
+	return content[:end]
+}
+
+// findCaseInsensitiveFile returns the on-disk path of the regular file in dir
+// whose name matches basename case-insensitively, or "" when absent.
+func findCaseInsensitiveFile(dir, basename string) string {
+	entries, err := os.ReadDir(dir)
+	if err != nil {
+		return ""
+	}
+	for _, e := range entries {
+		if !e.IsDir() && strings.EqualFold(e.Name(), basename) {
+			return filepath.Join(dir, e.Name())
+		}
+	}
+	return ""
+}
+
 // projectGuidelineDirs returns the directory chain from gitRoot to cwd
 // (inclusive), in root-to-leaf order. If gitRoot is empty or unreachable
 // from cwd, the chain collapses to [cwd].
@@ -378,23 +485,16 @@ func projectGuidelineLabel(match, gitRoot string) string {
 // Returns "" when nothing matches.
 func findProjectContextFile(dir string) string {
 	for _, name := range projectContextFiles {
-		baseLower := strings.ToLower(filepath.Base(name))
 		// Walk to the file's parent through dir with case-insensitive segment
-		// matching, then find a regular file with the same lowercased
-		// basename. This works on both case-sensitive and case-insensitive
-		// filesystems and always returns the on-disk filename.
+		// matching, then find a regular file with the same basename. This works
+		// on both case-sensitive and case-insensitive filesystems and always
+		// returns the on-disk filename.
 		parent, ok := resolveDirCaseInsensitive(filepath.Dir(filepath.Join(dir, name)), dir)
 		if !ok {
 			continue
 		}
-		entries, err := os.ReadDir(parent)
-		if err != nil {
-			continue
-		}
-		for _, e := range entries {
-			if !e.IsDir() && strings.ToLower(e.Name()) == baseLower {
-				return filepath.Join(parent, e.Name())
-			}
+		if match := findCaseInsensitiveFile(parent, filepath.Base(name)); match != "" {
+			return match
 		}
 	}
 	return ""

internal/agent/system_prompt.md

@@ -63,6 +63,10 @@ work.
   targeted diffs. Match the existing indentation, imports, and idioms. Match the
   file's comment density: do not add explanatory comments unless the user asks or
   the code is already comment-dense.
+- Solve the problem as posed, not a more general version of it. Add no
+  speculative abstraction, configurability, or handling for cases that cannot
+  occur, and nothing the user did not ask for. A small diff can still be
+  over-built; if a 200-line solution could be 50, rewrite it.
 - Preserve behavior you were not asked to change. Do not delete or rewrite code
   you did not author unless the task requires it; if you must, say so.
 

internal/agent/system_prompt_test.go

@@ -9,6 +9,25 @@ import (
 	"github.com/Gitlawb/zero/internal/sandbox"
 )
 
+// TestMain points userConfigDirForPrompt at an empty, package-wide temp
+// directory for every test in this package. Without this, buildSystemPrompt
+// falls back to the real config.UserConfigDir, so any developer with a
+// personal ~/.config/zero/ZERO.md would get its content folded into
+// otherwise-unrelated prompt assertions, making test runs non-deterministic
+// across contributor machines. Tests that specifically exercise user
+// guidelines stub userConfigDirForPrompt themselves via
+// withSystemPromptTestUserConfigDir(Func) and restore this default on cleanup.
+func TestMain(m *testing.M) {
+	dir, err := os.MkdirTemp("", "zero-agent-tests-*")
+	if err != nil {
+		panic(err)
+	}
+	userConfigDirForPrompt = func() (string, error) { return dir, nil }
+	code := m.Run()
+	os.RemoveAll(dir)
+	os.Exit(code)
+}
+
 func TestCoreSystemPromptIncludesCodingQualityRules(t *testing.T) {
 	prompt := strings.ToLower(buildSystemPrompt(Options{}))
 
@@ -128,6 +147,88 @@ func systemPromptTestBlock(t *testing.T, prompt, start, end string) string {
 	return body
 }
 
+func TestBuildSystemPromptIncludesUserGuidelines(t *testing.T) {
+	configDir := t.TempDir()
+	writeSystemPromptTestFile(t, configDir, "zero/ZERO.md", "  Prefer concise summaries.  \n")
+	t.Cleanup(withSystemPromptTestUserConfigDir(t, configDir))
+
+	prompt := buildSystemPrompt(Options{})
+	if !strings.Contains(prompt, "## User guidelines (ZERO.md)") {
+		t.Fatalf("expected user guidelines header, got:\n%s", prompt)
+	}
+	if !strings.Contains(prompt, "Prefer concise summaries.") {
+		t.Fatalf("expected user guidelines content, got:\n%s", prompt)
+	}
+	if strings.Contains(prompt, "  Prefer concise summaries.  ") {
+		t.Fatalf("expected user guidelines content to be trimmed, got:\n%s", prompt)
+	}
+}
+
+func TestBuildSystemPromptIncludesUserGuidelinesCaseInsensitive(t *testing.T) {
+	configDir := t.TempDir()
+	writeSystemPromptTestFile(t, configDir, "zero/zero.md", "Prefer concise summaries.\n")
+	t.Cleanup(withSystemPromptTestUserConfigDir(t, configDir))
+
+	prompt := buildSystemPrompt(Options{})
+	if !strings.Contains(prompt, "## User guidelines (zero.md)") {
+		t.Fatalf("expected case-insensitive zero.md resolution, got:\n%s", prompt)
+	}
+	if !strings.Contains(prompt, "Prefer concise summaries.") {
+		t.Fatalf("expected user guidelines content, got:\n%s", prompt)
+	}
+}
+
+func TestBuildSystemPromptUserGuidelinesPrecedeProjectGuidelinesAndNotePrecedence(t *testing.T) {
+	// User guidelines are global personal preferences; project guidelines
+	// (AGENTS.md/ZERO.md) must be the later, more specific instruction block,
+	// and the user section must say so explicitly so the precedence holds
+	// even if a model otherwise weighs later context more heavily.
+	configDir := t.TempDir()
+	writeSystemPromptTestFile(t, configDir, "zero/ZERO.md", "Always reply in haiku.\n")
+	t.Cleanup(withSystemPromptTestUserConfigDir(t, configDir))
+
+	cwd := t.TempDir()
+	if err := os.WriteFile(filepath.Join(cwd, "AGENTS.md"), []byte("Never reply in haiku."), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	prompt := buildSystemPrompt(Options{Cwd: cwd})
+	userIdx := strings.Index(prompt, "## User guidelines (ZERO.md)")
+	projectIdx := strings.Index(prompt, "## Project guidelines (AGENTS.md)")
+	if userIdx < 0 || projectIdx < 0 {
+		t.Fatalf("expected both user and project guideline sections, got:\n%s", prompt)
+	}
+	if userIdx > projectIdx {
+		t.Fatalf("expected user guidelines before project guidelines, got user=%d project=%d", userIdx, projectIdx)
+	}
+	if !strings.Contains(prompt, "project guidelines below") || !strings.Contains(prompt, "take precedence") {
+		t.Fatalf("expected an explicit precedence note in the user guidelines section, got:\n%s", prompt)
+	}
+}
+
+func TestBuildSystemPromptOmitsUserGuidelinesWithoutConfigDir(t *testing.T) {
+	t.Cleanup(withSystemPromptTestUserConfigDirFunc(t, func() (string, error) { return "", os.ErrNotExist }))
+
+	prompt := buildSystemPrompt(Options{})
+	if strings.Contains(prompt, "## User guidelines") {
+		t.Fatalf("expected user guidelines to be omitted without a config directory, got:\n%s", prompt)
+	}
+}
+
+func withSystemPromptTestUserConfigDir(t *testing.T, dir string) func() {
+	t.Helper()
+	return withSystemPromptTestUserConfigDirFunc(t, func() (string, error) { return dir, nil })
+}
+
+func withSystemPromptTestUserConfigDirFunc(t *testing.T, fn func() (string, error)) func() {
+	t.Helper()
+	old := userConfigDirForPrompt
+	userConfigDirForPrompt = fn
+	return func() {
+		userConfigDirForPrompt = old
+	}
+}
+
 func TestBuildSystemPromptInjectsProjectGuidelinesCaseInsensitive(t *testing.T) {
 	// Git tracks AGENTS.md on a case-sensitive filesystem; the
 	// loader must still resolve it when the cwd lookup uses lowercase.
@@ -257,6 +358,30 @@ func TestBuildSystemPromptProjectGuidelinesTruncatesAtTotalCap(t *testing.T) {
 	}
 }
 
+func TestTruncateGuidelineContentStaysWithinLimit(t *testing.T) {
+	limit := 64
+	content := strings.Repeat("x", limit+1)
+	got := truncateGuidelineContent(content, limit)
+	if len(got) > limit {
+		t.Fatalf("truncated result is %d bytes, want at most %d", len(got), limit)
+	}
+	if !strings.Contains(got, "truncated") {
+		t.Fatalf("expected truncation marker, got %q", got)
+	}
+}
+
+func TestTruncateGuidelineContentStaysWithinLimitForSmallBudgets(t *testing.T) {
+	// Limits at or below the truncation marker's own length can't fit the
+	// marker at all; the helper must still never exceed the requested limit.
+	content := strings.Repeat("x", 100)
+	for _, limit := range []int{0, 1, len(truncationMarker) - 1, len(truncationMarker), len(truncationMarker) + 1} {
+		got := truncateGuidelineContent(content, limit)
+		if len(got) > limit {
+			t.Fatalf("truncateGuidelineContent(_, %d) = %q (%d bytes), want at most %d bytes", limit, got, len(got), limit)
+		}
+	}
+}
+
 func TestProjectGuidelineDirsOrdersRootToLeaf(t *testing.T) {
 	root := filepath.Join("r")
 	leaf := filepath.Join(root, "a", "b")

internal/agent/types.go

@@ -285,6 +285,14 @@ type Options struct {
 	// ceiling and the autonomy gate. nil disables it entirely (the loop is
 	// byte-identical to before). One instance per run — it holds attempt state.
 	SelfCorrect *SelfCorrector
+	// FileDiagnostics, when set, checks files changed by mutating tools for
+	// error-severity language diagnostics IN THE BACKGROUND and appends any
+	// errors as a nudge before the model's next request — the model still sees
+	// an error it introduced at its next decision point, but no tool call ever
+	// blocks on the language server (the old inline path stalled every edit on
+	// a ≥300ms debounce, 10s cap). Build one with NewFileDiagnostics. nil
+	// disables post-edit diagnostics.
+	FileDiagnostics func(ctx context.Context, absPath string) string
 
 	// RequireCompletionSignal gates run completion for HEADLESS exec. Without it,
 	// any assistant turn that produces text but no tool call is accepted as the

internal/cli/app.go

@@ -12,6 +12,7 @@ import (
 	"os/exec"
 	"path/filepath"
 	"strings"
+	"sync"
 	"time"
 
 	"github.com/Gitlawb/zero/internal/agent"
@@ -19,6 +20,7 @@ import (
 	"github.com/Gitlawb/zero/internal/hooks"
 	"github.com/Gitlawb/zero/internal/localcontrol"
 	"github.com/Gitlawb/zero/internal/mcp"
+	"github.com/Gitlawb/zero/internal/modelregistry"
 	"github.com/Gitlawb/zero/internal/observability"
 	"github.com/Gitlawb/zero/internal/plugins"
 	"github.com/Gitlawb/zero/internal/providerhealth"
@@ -77,6 +79,8 @@ type appDeps struct {
 	runAgentEval           func(context.Context, agentEvalOptions) (agentEvalReport, error)
 	inspectChanges         func(context.Context, zerogit.InspectOptions) (zerogit.ChangeSummary, error)
 	commitChanges          func(context.Context, zerogit.CommitOptions) (zerogit.CommitResult, error)
+	pushChanges            func(context.Context, zerogit.PushOptions) (zerogit.PushResult, error)
+	createPR               func(context.Context, zerogit.PROptions) (zerogit.PRResult, error)
 	runTUI                 func(context.Context, tui.Options) int
 	runEditor              func(string) error
 	checkUpdate            func(context.Context, update.Options) (update.Result, error)
@@ -178,6 +182,8 @@ func defaultAppDeps() appDeps {
 		runAgentEval:     defaultRunAgentEval,
 		inspectChanges:   zerogit.Inspect,
 		commitChanges:    zerogit.Commit,
+		pushChanges:      zerogit.Push,
+		createPR:         zerogit.CreatePR,
 		runTUI:           tui.Run,
 		runEditor:        openEditor,
 		checkUpdate:      update.Check,
@@ -229,6 +235,12 @@ func runWithDeps(args []string, stdout io.Writer, stderr io.Writer, deps appDeps
 	defer observability.Recover(observability.DefaultCrashDir(), "cli", stderr, &exitCode)
 	deps = fillAppDeps(deps)
 
+	// CLI runs opt into the models.dev overlay (cached live context limits and
+	// pricing on top of the curated catalog). Explicitly enabled here — and only
+	// here — so library consumers and hermetic tests are never perturbed by a
+	// cache file on the machine. The refresh itself is fired in exec/TUI startup.
+	modelregistry.EnableModelsDevOverlay()
+
 	addDirs, args, err := splitLeadingAddDirFlags(args)
 	if err != nil {
 		return writeAppError(stderr, err.Error(), 1)
@@ -510,6 +522,12 @@ func fillAppDeps(deps appDeps) appDeps {
 	if deps.commitChanges == nil {
 		deps.commitChanges = defaults.commitChanges
 	}
+	if deps.pushChanges == nil {
+		deps.pushChanges = defaults.pushChanges
+	}
+	if deps.createPR == nil {
+		deps.createPR = defaults.createPR
+	}
 	if deps.runTUI == nil {
 		deps.runTUI = defaults.runTUI
 	}
@@ -544,6 +562,11 @@ func runInteractiveTUI(stderr io.Writer, deps appDeps, permissionMode agent.Perm
 }
 
 func runInteractiveTUIWithSetup(stderr io.Writer, deps appDeps, permissionMode agent.PermissionMode, addDirs []string, theme string, forceSetup bool) int {
+	// Refresh the models.dev pricing/limits cache in the background when stale;
+	// the overlay is read at registry construction from the cache file, so this
+	// benefits the next run and never blocks or fails this one.
+	go func() { _ = modelregistry.RefreshModelsDevCache(context.Background()) }()
+
 	workspaceRoot, err := deps.getwd()
 	if err != nil {
 		return writeAppError(stderr, "failed to resolve workspace: "+err.Error(), 1)
@@ -760,6 +783,14 @@ func runInteractiveTUIWithSetup(stderr io.Writer, deps appDeps, permissionMode a
 			Specialists:    specialistRuntime.specialists,
 			Skills:         pluginActivation.skillInfos(deps.skillsDir()),
 		},
+		// LoadSkills backs /skills and direct /<skill-name> invocation in the TUI.
+		// It resolves against the same merged set (default dir + plugin skill
+		// roots) as the skill tool and the system-prompt list, re-read per use so
+		// newly installed skills work without a restart.
+		LoadSkills: cachedSkillsLoader(func() []skills.Skill {
+			merged, _ := plugins.MergedSkillsLoaded(deps.skillsDir(), pluginActivation.skillRoots)
+			return merged
+		}),
 		PermissionMode: permissionMode,
 		Notify:         resolved.Notify,
 		KeyBindings:    resolved.KeyBindings,
@@ -1255,3 +1286,26 @@ Flags:
 `)
 	return err
 }
+
+// cachedSkillsLoader memoizes a skills loader for a short interval. The TUI
+// calls the loader from autocomplete on every "/x" keystroke; without a cache
+// each keystroke would re-read every SKILL.md body. Two seconds is fresh enough
+// that a newly installed skill still shows up "immediately" while typing.
+func cachedSkillsLoader(load func() []skills.Skill) func() []skills.Skill {
+	var (
+		mu     sync.Mutex
+		at     time.Time
+		cached []skills.Skill
+	)
+	const ttl = 2 * time.Second
+	return func() []skills.Skill {
+		mu.Lock()
+		defer mu.Unlock()
+		if cached != nil && time.Since(at) < ttl {
+			return cached
+		}
+		cached = load()
+		at = time.Now()
+		return cached
+	}
+}

internal/cli/exec.go

@@ -136,6 +136,11 @@ func runExec(args []string, stdout io.Writer, stderr io.Writer, deps appDeps) in
 		return exitSuccess
 	}
 
+	// Refresh the models.dev pricing/limits cache in the background when stale;
+	// the overlay is read at registry construction from the cache file, so this
+	// benefits the next run and never blocks or fails this one.
+	go func() { _ = modelregistry.RefreshModelsDevCache(context.Background()) }()
+
 	// A mode seeds model/effort/max-turns/tool filters as a preset. Expand it up
 	// front — before tool-filter validation and the --list-tools branch — so a
 	// mode-injected tool filter is validated and reflected in --list-tools, and a
@@ -500,9 +505,10 @@ func runExec(args []string, stdout io.Writer, stderr io.Writer, deps appDeps) in
 	// by default the corrector is nil, leaving the agent loop byte-identical. When
 	// on we verify with both the workspace test plan and LSP diagnostics over the
 	// changed files; the autonomy gate inside the corrector still decides whether
-	// failures auto-fix or just report. lspShutdown tears down any language-server
-	// sessions the LSP half spawned (no-op when self-correct is off).
-	selfCorrector, lspShutdown := newExecSelfCorrector(options.selfCorrect, workspaceRoot, options.autonomy)
+	// failures auto-fix or just report. fileDiagnostics (always on, lazy) gives
+	// edit_file/write_file inline error diagnostics for the file they just wrote.
+	// lspShutdown tears down any language-server sessions either half spawned.
+	selfCorrector, fileDiagnostics, lspShutdown := newExecSelfCorrector(options.selfCorrect, workspaceRoot, options.autonomy)
 	defer lspShutdown()
 	result, err := agent.Run(runCtx, agentPrompt, provider, agent.Options{
 		MaxTurns:         resolved.MaxTurns,
@@ -526,6 +532,7 @@ func runExec(args []string, stdout io.Writer, stderr io.Writer, deps appDeps) in
 		PermissionMode:   permissionMode,
 		Autonomy:         options.autonomy,
 		SelfCorrect:      selfCorrector,
+		FileDiagnostics:  fileDiagnostics,
 		// Headless exec: don't accept a no-tool-call turn as "done" while work
 		// clearly remains (pending plan items / a mid-step continuation cue) —
 		// nudge to continue, and finalize as INCOMPLETE rather than false success
@@ -537,6 +544,7 @@ func runExec(args []string, stdout io.Writer, stderr io.Writer, deps appDeps) in
 		EnabledTools:            options.enabledTools,
 		DisabledTools:           options.disabledTools,
 		OnText:                  writer.text,
+		OnReasoning:             writer.reasoning,
 		OnToolCall: func(call agent.ToolCall) {
 			writer.toolCall(call, registry)
 			sessionRecorder.append(sessions.EventToolCall, map[string]any{
@@ -701,23 +709,29 @@ func runExec(args []string, stdout io.Writer, stderr io.Writer, deps appDeps) in
 // unless a changed file's language actually has one installed on PATH. The
 // returned cleanup shuts that manager down (terminating any spawned server
 // sessions); it is a no-op when self-correct is off.
-func newExecSelfCorrector(enabled bool, workspaceRoot string, autonomy string) (*agent.SelfCorrector, func()) {
-	if !enabled {
-		return nil, func() {}
-	}
+func newExecSelfCorrector(enabled bool, workspaceRoot string, autonomy string) (*agent.SelfCorrector, func(context.Context, string) string, func()) {
+	// The manager is created regardless of --self-correct: it also backs the
+	// always-on background post-edit diagnostics (agent.NewFileDiagnostics,
+	// collected off the tool-call path by the loop), and it stays lazy — no
+	// language server is spawned unless an edited file's language actually has
+	// one installed on PATH.
 	manager := lsp.NewManager(workspaceRoot)
+	cleanup := func() {
+		ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+		defer cancel()
+		_ = manager.Shutdown(ctx)
+	}
+	fileDiagnostics := agent.NewFileDiagnostics(manager, workspaceRoot)
+	if !enabled {
+		return nil, fileDiagnostics, cleanup
+	}
 	corrector := agent.NewSelfCorrector(workspaceRoot, agent.NewLSPDiagnosticsChecker(manager), agent.NewProjectVerifier(workspaceRoot), agent.SelfCorrectConfig{
 		Enabled:      true,
 		IncludeTests: true,
 		IncludeLSP:   true,
 		Autonomy:     autonomy,
 	})
-	cleanup := func() {
-		ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
-		defer cancel()
-		_ = manager.Shutdown(ctx)
-	}
-	return corrector, cleanup
+	return corrector, fileDiagnostics, cleanup
 }
 
 func deferredEligibleCount(registry *tools.Registry, permissionMode agent.PermissionMode, enabledTools []string, disabledTools []string) int {

internal/cli/exec_protocol_test.go

@@ -529,6 +529,45 @@ func TestRunExecStreamJSONRunStartUsesResolvedAPIModel(t *testing.T) {
 	}
 }
 
+func TestRunExecStreamJSONEmitsReasoningEvents(t *testing.T) {
+	t.Setenv("XDG_DATA_HOME", t.TempDir())
+	cwd := t.TempDir()
+	var stdout bytes.Buffer
+	var stderr bytes.Buffer
+
+	exitCode := runWithDeps([]string{"exec", "--output-format", "stream-json", "think then answer"}, &stdout, &stderr, appDeps{
+		getwd: func() (string, error) {
+			return cwd, nil
+		},
+		resolveConfig: func(_ string, _ config.Overrides) (config.ResolvedConfig, error) {
+			return execResolvedConfig(), nil
+		},
+		newProvider: func(config.ProviderProfile) (zeroruntime.Provider, error) {
+			return reasoningExecProvider{}, nil
+		},
+	})
+
+	if exitCode != exitSuccess {
+		t.Fatalf("expected exit code %d, got %d: %s", exitSuccess, exitCode, stderr.String())
+	}
+	events := decodeJSONLines(t, stdout.String())
+	reasoningIdx, reasoning := findJSONEventIndex(t, events, "reasoning")
+	if reasoning["delta"] != "Thinking. " {
+		t.Fatalf("unexpected reasoning event: %#v", reasoning)
+	}
+	textIdx, text := findJSONEventIndex(t, events, "text")
+	if text["delta"] != "done" {
+		t.Fatalf("unexpected text event: %#v", text)
+	}
+	if reasoningIdx >= textIdx {
+		t.Fatalf("expected reasoning event before text event, got indices %d and %d", reasoningIdx, textIdx)
+	}
+	final := findJSONEvent(t, events, "final")
+	if final["text"] != "done" {
+		t.Fatalf("reasoning must not be folded into final answer: %#v", final)
+	}
+}
+
 func TestExecEventWriterTruncatesStreamJSONToolResults(t *testing.T) {
 	var stdout bytes.Buffer
 	writer := execEventWriter{
@@ -751,6 +790,17 @@ type toolCallingExecProvider struct {
 	answer     string
 }
 
+type reasoningExecProvider struct{}
+
+func (reasoningExecProvider) StreamCompletion(context.Context, zeroruntime.CompletionRequest) (<-chan zeroruntime.StreamEvent, error) {
+	ch := make(chan zeroruntime.StreamEvent, 3)
+	ch <- zeroruntime.StreamEvent{Type: zeroruntime.StreamEventReasoning, Content: "Thinking. "}
+	ch <- zeroruntime.StreamEvent{Type: zeroruntime.StreamEventText, Content: "done"}
+	ch <- zeroruntime.StreamEvent{Type: zeroruntime.StreamEventDone}
+	close(ch)
+	return ch, nil
+}
+
 func (provider toolCallingExecProvider) StreamCompletion(ctx context.Context, request zeroruntime.CompletionRequest) (<-chan zeroruntime.StreamEvent, error) {
 	for _, message := range request.Messages {
 		if message.Role == zeroruntime.MessageRoleTool {
@@ -777,13 +827,19 @@ func (provider toolCallingExecProvider) StreamCompletion(ctx context.Context, re
 
 func findJSONEvent(t *testing.T, events []map[string]any, eventType string) map[string]any {
 	t.Helper()
-	for _, event := range events {
+	_, event := findJSONEventIndex(t, events, eventType)
+	return event
+}
+
+func findJSONEventIndex(t *testing.T, events []map[string]any, eventType string) (int, map[string]any) {
+	t.Helper()
+	for idx, event := range events {
 		if event["type"] == eventType {
-			return event
+			return idx, event
 		}
 	}
 	t.Fatalf("event %q not found in %#v", eventType, events)
-	return nil
+	return -1, nil
 }
 
 func findSessionEvent(t *testing.T, events []sessions.Event, eventType sessions.EventType) sessions.Event {

internal/cli/exec_writer.go

@@ -80,6 +80,16 @@ func (writer *execEventWriter) text(delta string) {
 	writer.writeStdout(delta)
 }
 
+func (writer *execEventWriter) reasoning(delta string) {
+	if writer.format == execOutputJSON {
+		writer.writeJSON(map[string]any{"type": "reasoning", "delta": delta})
+		return
+	}
+	if writer.format == execOutputStreamJSON {
+		writer.writeStreamJSON(streamjson.Event{Type: streamjson.EventReasoning, RunID: writer.runID, Delta: delta})
+	}
+}
+
 func (writer *execEventWriter) toolCall(call agent.ToolCall, registry *tools.Registry) {
 	if writer.format == execOutputJSON {
 		writer.writeJSON(map[string]any{

internal/cli/repoinfo.go

@@ -88,6 +88,15 @@ func formatRepoInfo(info repoinfo.Info) string {
 	fmt.Fprintf(&b, "Files:        %d\n", info.FileCount)
 	fmt.Fprintf(&b, "Directories:  %d (max depth %d)\n", info.DirectoryCount, info.MaxDepth)
 	fmt.Fprintf(&b, "Est. LOC:     %d\n", info.LOCEstimate)
+	if info.CommitCount != nil {
+		fmt.Fprintf(&b, "Commits:      %d\n", *info.CommitCount)
+	}
+	if info.BranchCount != nil {
+		fmt.Fprintf(&b, "Branches:     %d\n", *info.BranchCount)
+	}
+	if info.TagCount != nil {
+		fmt.Fprintf(&b, "Tags:         %d\n", *info.TagCount)
+	}
 	if info.PrimaryLanguage != "" {
 		fmt.Fprintf(&b, "Primary lang: %s\n", info.PrimaryLanguage)
 	}

internal/cli/repoinfo_test.go

@@ -33,7 +33,7 @@ func TestFormatRepoInfoText(t *testing.T) {
 	contrib := 3
 	info := repoInfoTestData(age, contrib)
 	out := formatRepoInfo(info)
-	for _, want := range []string{"Files", "TypeScript", "main", "GitHub Actions", "42", "3"} {
+	for _, want := range []string{"Files", "TypeScript", "main", "GitHub Actions", "42", "3", "Commits:      100", "Branches:     5", "Tags:         10"} {
 		if !strings.Contains(out, want) {
 			t.Fatalf("text output missing %q:\n%s", want, out)
 		}
@@ -113,12 +113,16 @@ func initTempGitRepo(t *testing.T) string {
 }
 
 func repoInfoTestData(age, contrib int) repoinfo.Info {
+	commits := 100
+	branches := 5
+	tags := 10
 	return repoinfo.Info{
 		FileCount: 10, DirectoryCount: 3, MaxDepth: 2, LOCEstimate: 500,
 		Languages:       []repoinfo.LangStat{{Name: "TypeScript", LOCEstimate: 400, FileCount: 4}},
 		PrimaryLanguage: "TypeScript", LanguageCount: 1,
 		WorkspaceType: "none", CICD: []string{"GitHub Actions"},
 		HasGit: true, Branch: "main", AgeDays: &age, Contributors90d: &contrib,
+		CommitCount: &commits, BranchCount: &branches, TagCount: &tags,
 	}
 }
 

internal/cli/workflows.go

@@ -41,6 +41,13 @@ type changesCommandOptions struct {
 	hasMessage   bool
 	dryRun       bool
 	maxDiffBytes int
+	remote       string
+	force        bool
+	title        string
+	body         string
+	fill         bool
+	draft        bool
+	yes          bool
 	auto         bool
 }
 
@@ -266,6 +273,10 @@ func runChanges(args []string, stdout io.Writer, stderr io.Writer, deps appDeps)
 			return exitCrash
 		}
 		return exitSuccess
+	case "push":
+		return runChangesPush(args, stdout, stderr, deps)
+	case "pr", "pull-request":
+		return runChangesPR(args, stdout, stderr, deps)
 	default:
 		return writeExecUsageError(stderr, fmt.Sprintf("unknown changes command %q", command))
 	}
@@ -465,6 +476,41 @@ func parseChangesArgs(args []string, command string) (changesCommandOptions, boo
 				return options, false, err
 			}
 			options.maxDiffBytes = maxDiffBytes
+		case arg == "--remote":
+			value, next, err := nextFlagValue(args, index, arg)
+			if err != nil {
+				return options, false, err
+			}
+			options.remote = strings.TrimSpace(value)
+			index = next
+		case strings.HasPrefix(arg, "--remote="):
+			options.remote = strings.TrimSpace(strings.TrimPrefix(arg, "--remote="))
+		case arg == "--force":
+			options.force = true
+		case arg == "--title":
+			value, next, err := nextFlagValue(args, index, arg)
+			if err != nil {
+				return options, false, err
+			}
+			options.title = value
+			index = next
+		case strings.HasPrefix(arg, "--title="):
+			options.title = strings.TrimPrefix(arg, "--title=")
+		case arg == "--body":
+			value, next, err := nextFlagValue(args, index, arg)
+			if err != nil {
+				return options, false, err
+			}
+			options.body = value
+			index = next
+		case strings.HasPrefix(arg, "--body="):
+			options.body = strings.TrimPrefix(arg, "--body=")
+		case arg == "--fill":
+			options.fill = true
+		case arg == "--draft":
+			options.draft = true
+		case arg == "--yes":
+			options.yes = true
 		case arg == "-a" || arg == "--auto":
 			options.auto = true
 		case strings.HasPrefix(arg, "-"):
@@ -473,15 +519,30 @@ func parseChangesArgs(args []string, command string) (changesCommandOptions, boo
 			return options, false, execUsageError{fmt.Sprintf("unexpected changes argument %q", arg)}
 		}
 	}
+	if command != "commit" && options.message != "" {
+		return options, false, execUsageError{"--message is only valid with `zero changes commit`"}
+	}
 	if command != "commit" && (options.hasMessage || options.dryRun || options.auto) {
 		return options, false, execUsageError{"--message, --dry-run, and --auto are only valid with `zero changes commit`"}
 	}
 	if command == "commit" && options.hasMessage && options.auto {
 		return options, false, execUsageError{"cannot specify both --message and --auto"}
 	}
-	if command == "commit" && options.baseRef != "" {
+	if command != "commit" && command != "push" && options.dryRun {
+		return options, false, execUsageError{"--dry-run is only valid with commit or push"}
+	}
+	if command != "inspect" && options.baseRef != "" {
 		return options, false, execUsageError{"--base is only valid with `zero changes inspect`"}
 	}
+	if command != "push" && command != "pr" && (options.remote != "" || options.force) {
+		return options, false, execUsageError{"--remote and --force are only valid with push or pr"}
+	}
+	if command != "pr" && (options.title != "" || options.body != "" || options.fill || options.draft) {
+		return options, false, execUsageError{"--title, --body, --fill, and --draft are only valid with `zero changes pr`"}
+	}
+	if command != "push" && command != "pr" && options.yes {
+		return options, false, execUsageError{"--yes is only valid with push or pr"}
+	}
 	return options, false, nil
 }
 
@@ -759,14 +820,24 @@ func writeChangesHelp(w io.Writer) error {
 	_, err := fmt.Fprint(w, `Usage:
   zero changes inspect [flags]
   zero changes commit [flags]
+  zero changes push [flags]
+  zero changes pr [flags]
 
-Inspects local git changes and optionally creates a commit.
+Inspects, commits, pushes, and creates pull requests for local git changes.
 
 Flags:
   -C, --cwd <path>        Workspace directory
       --base <ref>        Diff against <ref>...HEAD instead of the working tree
       --diff-bytes <n>    Maximum diff bytes to include
   -m, --message <text>    Commit message for `+"`zero changes commit`"+`
+      --dry-run           Preview commit metadata / push without mutating git state
+      --remote <name>     Remote to push to (defaults to upstream tracked branch or origin)
+      --force             Use force-with-lease when pushing
+      --title <text>      PR title
+      --body <text>       PR body
+      --fill              Automatically populate PR title and body from commits
+      --draft             Create PR as a draft
+      --yes               Confirm pushing to a default/protected branch
   -a, --auto              Auto-generate commit message using LLM (use --dry-run to preview)
       --dry-run           Preview commit metadata without mutating git state
       --json              Print JSON output
@@ -775,6 +846,128 @@ Flags:
 	return err
 }
 
+func runChangesPush(args []string, stdout io.Writer, stderr io.Writer, deps appDeps) int {
+	options, help, err := parseChangesArgs(args, "push")
+	if err != nil {
+		return writeExecUsageError(stderr, err.Error())
+	}
+	if help {
+		if err := writeChangesHelp(stdout); err != nil {
+			return exitCrash
+		}
+		return exitSuccess
+	}
+	workspaceRoot, err := resolveWorkspaceRoot(options.cwd, deps)
+	if err != nil {
+		return writeExecUsageError(stderr, err.Error())
+	}
+
+	result, err := deps.pushChanges(context.Background(), zerogit.PushOptions{
+		Cwd:                    workspaceRoot,
+		Remote:                 options.remote,
+		Force:                  options.force,
+		DryRun:                 options.dryRun,
+		AllowPushDefaultBranch: options.yes,
+	})
+	if err != nil {
+		return writeExecUsageError(stderr, err.Error())
+	}
+
+	if options.json {
+		if err := writePrettyJSON(stdout, result); err != nil {
+			return exitCrash
+		}
+		return exitSuccess
+	}
+
+	dryRunStr := ""
+	if options.dryRun {
+		dryRunStr = " (dry run)"
+	}
+	if _, err := fmt.Fprintf(stdout, "Pushed branch %s to remote %s%s\n", result.Branch, result.Remote, dryRunStr); err != nil {
+		return exitCrash
+	}
+	if result.Output != "" {
+		if _, err := fmt.Fprintln(stdout, result.Output); err != nil {
+			return exitCrash
+		}
+	}
+	return exitSuccess
+}
+
+func runChangesPR(args []string, stdout io.Writer, stderr io.Writer, deps appDeps) int {
+	options, help, err := parseChangesArgs(args, "pr")
+	if err != nil {
+		return writeExecUsageError(stderr, err.Error())
+	}
+	if help {
+		if err := writeChangesHelp(stdout); err != nil {
+			return exitCrash
+		}
+		return exitSuccess
+	}
+	if !options.fill && options.title == "" {
+		return writeExecUsageError(stderr, "must provide either --fill or --title to run non-interactively")
+	}
+	workspaceRoot, err := resolveWorkspaceRoot(options.cwd, deps)
+	if err != nil {
+		return writeExecUsageError(stderr, err.Error())
+	}
+
+	if !options.json {
+		if _, err := fmt.Fprintln(stdout, "Pushing current branch to set upstream..."); err != nil {
+			return exitCrash
+		}
+	}
+	pushResult, err := deps.pushChanges(context.Background(), zerogit.PushOptions{
+		Cwd:                    workspaceRoot,
+		Remote:                 options.remote,
+		Force:                  options.force,
+		AllowPushDefaultBranch: options.yes,
+	})
+	if err != nil {
+		return writeExecUsageError(stderr, fmt.Sprintf("auto-push failed: %v", err))
+	}
+	if !options.json {
+		if _, err := fmt.Fprintf(stdout, "Pushed branch %s to remote %s\n", pushResult.Branch, pushResult.Remote); err != nil {
+			return exitCrash
+		}
+	}
+
+	prResult, err := deps.createPR(context.Background(), zerogit.PROptions{
+		Cwd:   workspaceRoot,
+		Fill:  options.fill,
+		Draft: options.draft,
+		Title: options.title,
+		Body:  options.body,
+	})
+	if err != nil {
+		return writeExecUsageError(stderr, err.Error())
+	}
+
+	if options.json {
+		type prJSONResult struct {
+			Branch string `json:"branch"`
+			Remote string `json:"remote"`
+			Output string `json:"output"`
+		}
+		res := prJSONResult{
+			Branch: pushResult.Branch,
+			Remote: pushResult.Remote,
+			Output: strings.TrimSpace(prResult.Output),
+		}
+		if err := writePrettyJSON(stdout, res); err != nil {
+			return exitCrash
+		}
+		return exitSuccess
+	}
+
+	if _, err := fmt.Fprintln(stdout, strings.TrimSpace(prResult.Output)); err != nil {
+		return exitCrash
+	}
+	return exitSuccess
+}
+
 func generateAutoCommitMessage(ctx context.Context, provider zeroruntime.Provider, model string, summary zerogit.ChangeSummary) (string, error) {
 	var promptBuilder strings.Builder
 	promptBuilder.WriteString("Analyze the following git diff and generate a concise, conventional commit message.\n")

internal/doctor/hardening.go

@@ -131,29 +131,45 @@ func doctorSandboxPolicy(cfg config.SandboxConfig) sandbox.Policy {
 // lspServersCheck reports which language servers ZERO would use are present on
 // PATH. Missing servers are not a failure — ZERO degrades to text-only edits for
 // those languages — so the worst status is WARN, and each missing server gets an
-// actionable install command keyed by its binary name.
+// actionable install command keyed by its binary name. Only the tier-1 servers
+// (lsp.CoreServerBinaries) drive the status: the long-tail servers configured
+// for breadth are listed informationally under missingOptional, because warning
+// about a missing zls on a machine with no Zig code would be permanent noise.
 func lspServersCheck(lookup func(string) (string, error)) Check {
 	if lookup == nil {
 		lookup = exec.LookPath
 	}
+	core := map[string]bool{}
+	for _, binary := range lsp.CoreServerBinaries() {
+		core[binary] = true
+	}
 	present := map[string]any{}
 	missing := map[string]any{}
+	missingOptional := map[string]any{}
 	for _, binary := range lsp.ServerBinaries() {
 		if _, err := lookup(binary); err == nil {
 			present[binary] = "on PATH"
 			continue
 		}
-		missing[binary] = lspRemedy(binary)
+		if core[binary] {
+			missing[binary] = lspRemedy(binary)
+		} else {
+			missingOptional[binary] = lspRemedy(binary)
+		}
+	}
+	details := map[string]any{"present": present}
+	if len(missingOptional) > 0 {
+		details["missingOptional"] = missingOptional
 	}
 	if len(missing) == 0 {
-		return check("lsp.servers", "LSP servers", StatusPass, "All known language servers are available on PATH.", map[string]any{
-			"present": present,
-		})
+		message := "All core language servers are available on PATH."
+		if len(missingOptional) > 0 {
+			message = fmt.Sprintf("All core language servers are available on PATH (%d optional server(s) not installed).", len(missingOptional))
+		}
+		return check("lsp.servers", "LSP servers", StatusPass, message, details)
 	}
-	return check("lsp.servers", "LSP servers", StatusWarn, fmt.Sprintf("%d language server(s) missing from PATH; affected files degrade to text-only edits.", len(missing)), map[string]any{
-		"present": present,
-		"missing": missing,
-	})
+	details["missing"] = missing
+	return check("lsp.servers", "LSP servers", StatusWarn, fmt.Sprintf("%d language server(s) missing from PATH; affected files degrade to text-only edits.", len(missing)), details)
 }
 
 // lspRemedy returns an actionable install command for a missing language-server

internal/lsp/registry.go

@@ -9,26 +9,134 @@ import (
 
 // serverCommands maps a file extension to the language-server command (argv) ZERO
 // will spawn. The first element is the binary looked up on PATH; missing binaries
-// are not an error, the agent just degrades to text-only for that file.
+// are not an error, the agent just degrades to text-only for that file. Every
+// command is the language's community-standard server invoked in stdio mode, so
+// a wrong guess can never spawn something surprising — only fail the PATH check.
 var serverCommands = map[string][]string{
-	".go":  {"gopls", "serve"},
-	".ts":  {"typescript-language-server", "--stdio"},
-	".tsx": {"typescript-language-server", "--stdio"},
-	".js":  {"typescript-language-server", "--stdio"},
-	".jsx": {"typescript-language-server", "--stdio"},
-	".py":  {"pyright-langserver", "--stdio"},
-	".rs":  {"rust-analyzer"},
+	".go":     {"gopls", "serve"},
+	".ts":     {"typescript-language-server", "--stdio"},
+	".tsx":    {"typescript-language-server", "--stdio"},
+	".mts":    {"typescript-language-server", "--stdio"},
+	".cts":    {"typescript-language-server", "--stdio"},
+	".js":     {"typescript-language-server", "--stdio"},
+	".jsx":    {"typescript-language-server", "--stdio"},
+	".mjs":    {"typescript-language-server", "--stdio"},
+	".cjs":    {"typescript-language-server", "--stdio"},
+	".py":     {"pyright-langserver", "--stdio"},
+	".pyi":    {"pyright-langserver", "--stdio"},
+	".rs":     {"rust-analyzer"},
+	".c":      {"clangd"},
+	".h":      {"clangd"},
+	".cpp":    {"clangd"},
+	".cc":     {"clangd"},
+	".cxx":    {"clangd"},
+	".hpp":    {"clangd"},
+	".java":   {"jdtls"},
+	".kt":     {"kotlin-language-server"},
+	".kts":    {"kotlin-language-server"},
+	".rb":     {"ruby-lsp"},
+	".php":    {"intelephense", "--stdio"},
+	".zig":    {"zls"},
+	".lua":    {"lua-language-server"},
+	".ex":     {"elixir-ls"},
+	".exs":    {"elixir-ls"},
+	".hs":     {"haskell-language-server-wrapper", "--lsp"},
+	".swift":  {"sourcekit-lsp"},
+	".ml":     {"ocamllsp"},
+	".mli":    {"ocamllsp"},
+	".scala":  {"metals"},
+	".clj":    {"clojure-lsp"},
+	".cljs":   {"clojure-lsp"},
+	".cljc":   {"clojure-lsp"},
+	".dart":   {"dart", "language-server"},
+	".gleam":  {"gleam", "lsp"},
+	".nix":    {"nixd"},
+	".tf":     {"terraform-ls", "serve"},
+	".sh":     {"bash-language-server", "start"},
+	".bash":   {"bash-language-server", "start"},
+	".yaml":   {"yaml-language-server", "--stdio"},
+	".yml":    {"yaml-language-server", "--stdio"},
+	".json":   {"vscode-json-language-server", "--stdio"},
+	".css":    {"vscode-css-language-server", "--stdio"},
+	".scss":   {"vscode-css-language-server", "--stdio"},
+	".less":   {"vscode-css-language-server", "--stdio"},
+	".html":   {"vscode-html-language-server", "--stdio"},
+	".svelte": {"svelteserver", "--stdio"},
+	".vue":    {"vue-language-server", "--stdio"},
+	".astro":  {"astro-ls", "--stdio"},
 }
 
 // languageIDs maps a file extension to the LSP languageId used in didOpen.
 var languageIDs = map[string]string{
-	".go":  "go",
-	".ts":  "typescript",
-	".tsx": "typescriptreact",
-	".js":  "javascript",
-	".jsx": "javascriptreact",
-	".py":  "python",
-	".rs":  "rust",
+	".go":     "go",
+	".ts":     "typescript",
+	".tsx":    "typescriptreact",
+	".mts":    "typescript",
+	".cts":    "typescript",
+	".js":     "javascript",
+	".jsx":    "javascriptreact",
+	".mjs":    "javascript",
+	".cjs":    "javascript",
+	".py":     "python",
+	".pyi":    "python",
+	".rs":     "rust",
+	".c":      "c",
+	".h":      "c",
+	".cpp":    "cpp",
+	".cc":     "cpp",
+	".cxx":    "cpp",
+	".hpp":    "cpp",
+	".java":   "java",
+	".kt":     "kotlin",
+	".kts":    "kotlin",
+	".rb":     "ruby",
+	".php":    "php",
+	".zig":    "zig",
+	".lua":    "lua",
+	".ex":     "elixir",
+	".exs":    "elixir",
+	".hs":     "haskell",
+	".swift":  "swift",
+	".ml":     "ocaml",
+	".mli":    "ocaml",
+	".scala":  "scala",
+	".clj":    "clojure",
+	".cljs":   "clojure",
+	".cljc":   "clojure",
+	".dart":   "dart",
+	".gleam":  "gleam",
+	".nix":    "nix",
+	".tf":     "terraform",
+	".sh":     "shellscript",
+	".bash":   "shellscript",
+	".yaml":   "yaml",
+	".yml":    "yaml",
+	".json":   "json",
+	".css":    "css",
+	".scss":   "scss",
+	".less":   "less",
+	".html":   "html",
+	".svelte": "svelte",
+	".vue":    "vue",
+	".astro":  "astro",
+}
+
+// coreServerBinaries are the tier-1 servers for the languages agents hit most;
+// `zero doctor` treats their absence as warn-worthy. The long tail of servers
+// configured above for breadth is reported informationally only — warning about
+// a missing zls on a machine with no Zig code would be permanent noise.
+var coreServerBinaries = []string{
+	"gopls",
+	"typescript-language-server",
+	"pyright-langserver",
+	"rust-analyzer",
+}
+
+// CoreServerBinaries returns the tier-1 server binaries (sorted copy).
+func CoreServerBinaries() []string {
+	binaries := append([]string(nil), coreServerBinaries...)
+	sort.Strings(binaries)
+	return binaries
 }
 
 // ServerBinaries returns the unique set of language-server binaries ZERO may

internal/modelregistry/catalog.go

@@ -63,6 +63,10 @@ func DefaultModelEntries() []ModelEntry {
 		googleModel("gemini-2.5-flash-lite", "Gemini 2.5 Flash-Lite", "gemini-2.5-flash-lite", ModelStatusActive, []string{"google:gemini-2.5-flash-lite", "gemini-flash-lite"}, ContextLimits{ContextWindow: 1_048_576, MaxOutputTokens: 65_536}, ModelCost{InputPerMillion: 0.1, CachedInputPerMillion: 0.01, OutputPerMillion: 0.4}, []ModelCapability{ModelCapabilityVision, ModelCapabilityJSONMode, ModelCapabilityReasoning, ModelCapabilityLongContext}, standardReasoningEfforts(), "Google low-cost Flash model for background routing and summaries."),
 	}
 	decorateModelDepth(entries)
+	// Overlay volatile facts (context limits, base pricing) from a cached
+	// models.dev snapshot when one is present and fresh — see modelsdev.go.
+	// Identity fields (ids, aliases, patterns, deprecations) stay curated.
+	entries = applyModelsDevOverrides(entries, cachedModelsDevProviders())
 	return cloneModelEntries(entries)
 }
 

internal/modelregistry/modelsdev.go

@@ -0,0 +1,269 @@
+package modelregistry
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+	"path/filepath"
+	"strings"
+	"sync"
+	"sync/atomic"
+	"time"
+)
+
+// Live models.dev overlay for the curated catalog. The hand-maintained
+// DefaultModelEntries list is the source of truth for identity — ids, aliases,
+// match patterns, deprecations, escalation targets — but its VOLATILE facts
+// (context window, max output tokens, per-million pricing) go stale between
+// releases. When a cached snapshot of https://models.dev/api.json is present,
+// those fields are refreshed from it at registry construction; everything else
+// stays curated. The overlay never adds models (an auto-added entry would lack
+// aliases, match patterns, and provider wiring) and never touches the network
+// on the registry hot path — fetching happens only in the explicit background
+// refresh, cached to disk with a TTL.
+
+const (
+	modelsDevDefaultURL = "https://models.dev/api.json"
+	// modelsDevRefreshAfter is how old the cache may get before a background
+	// refresh re-fetches it.
+	modelsDevRefreshAfter = 24 * time.Hour
+	// modelsDevMaxAge is the oldest cache still applied as an overlay. Beyond
+	// this the curated catalog (updated with the binary) is likely fresher than
+	// the snapshot, so a stale file is ignored rather than trusted.
+	modelsDevMaxAge      = 7 * 24 * time.Hour
+	modelsDevFetchLimit  = 32 << 20 // 32MiB guard on the response body
+	modelsDevFetchWindow = 15 * time.Second
+)
+
+// modelsDevModel is the subset of a models.dev model record the overlay uses.
+type modelsDevModel struct {
+	Limit struct {
+		Context int `json:"context"`
+		Output  int `json:"output"`
+	} `json:"limit"`
+	Cost struct {
+		Input      float64 `json:"input"`
+		Output     float64 `json:"output"`
+		CacheRead  float64 `json:"cache_read"`
+		CacheWrite float64 `json:"cache_write"`
+	} `json:"cost"`
+}
+
+// modelsDevProvider matches one provider object in api.json.
+type modelsDevProvider struct {
+	Models map[string]modelsDevModel `json:"models"`
+}
+
+// parseModelsDev decodes an api.json document into provider-slug -> api-model
+// -> record. api.json is a top-level object keyed by provider id.
+func parseModelsDev(data []byte) (map[string]map[string]modelsDevModel, error) {
+	var doc map[string]modelsDevProvider
+	if err := json.Unmarshal(data, &doc); err != nil {
+		return nil, fmt.Errorf("modelregistry: parse models.dev: %w", err)
+	}
+	providers := make(map[string]map[string]modelsDevModel, len(doc))
+	for slug, provider := range doc {
+		if len(provider.Models) == 0 {
+			continue
+		}
+		providers[slug] = provider.Models
+	}
+	if len(providers) == 0 {
+		return nil, fmt.Errorf("modelregistry: models.dev document has no providers")
+	}
+	return providers, nil
+}
+
+// modelsDevSlugs maps an entry's provider kind to the models.dev provider ids
+// worth checking. Only first-party slugs are used: router entries on models.dev
+// carry generic or stale numbers.
+func modelsDevSlugs(kind ProviderKind) []string {
+	switch kind {
+	case ProviderAnthropic:
+		return []string{"anthropic"}
+	case ProviderOpenAI:
+		return []string{"openai"}
+	case ProviderGoogle:
+		return []string{"google", "google-vertex"}
+	default:
+		return nil
+	}
+}
+
+// applyModelsDevOverrides refreshes each curated entry's context limits and
+// base pricing from the snapshot, when the snapshot knows the model. Tiered
+// pricing is never touched: models.dev has no tier data, and mixing a live
+// base rate with curated tiers would misprice the tier boundaries.
+func applyModelsDevOverrides(entries []ModelEntry, providers map[string]map[string]modelsDevModel) []ModelEntry {
+	if len(providers) == 0 {
+		return entries
+	}
+	for i := range entries {
+		entry := &entries[i]
+		var record modelsDevModel
+		found := false
+		for _, slug := range modelsDevSlugs(entry.Provider) {
+			if models, ok := providers[slug]; ok {
+				if candidate, ok := models[strings.TrimSpace(entry.APIModel)]; ok {
+					record = candidate
+					found = true
+					break
+				}
+			}
+		}
+		if !found {
+			continue
+		}
+		if record.Limit.Context > 0 {
+			entry.ContextLimits.ContextWindow = record.Limit.Context
+		}
+		if record.Limit.Output > 0 {
+			entry.ContextLimits.MaxOutputTokens = record.Limit.Output
+		}
+		if len(entry.Cost.Tiers) == 0 && record.Cost.Input > 0 && record.Cost.Output > 0 {
+			entry.Cost.InputPerMillion = record.Cost.Input
+			entry.Cost.OutputPerMillion = record.Cost.Output
+			if record.Cost.CacheRead > 0 {
+				entry.Cost.CachedInputPerMillion = record.Cost.CacheRead
+			}
+			if record.Cost.CacheWrite > 0 {
+				entry.Cost.CacheWritePerMillion = record.Cost.CacheWrite
+			}
+			entry.Cost.Source = "models.dev/api.json (cached)"
+		}
+	}
+	return entries
+}
+
+// modelsDevCachePath returns the on-disk cache location. ZERO_MODELS_CACHE_PATH
+// overrides it (used by tests and unusual setups).
+func modelsDevCachePath() (string, error) {
+	if override := strings.TrimSpace(os.Getenv("ZERO_MODELS_CACHE_PATH")); override != "" {
+		return override, nil
+	}
+	base, err := os.UserCacheDir()
+	if err != nil {
+		return "", err
+	}
+	return filepath.Join(base, "zero", "modelsdev.json"), nil
+}
+
+var (
+	modelsDevOnce    sync.Once
+	modelsDevCached  map[string]map[string]modelsDevModel
+	modelsDevEnabled atomic.Bool
+)
+
+// EnableModelsDevOverlay opts the process into applying the cached models.dev
+// snapshot on top of the curated catalog. The CLI entrypoint calls it; library
+// consumers and tests that never do get the curated catalog byte-identical to
+// before, so hermetic tests can't be perturbed by a cache file on the machine.
+// ZERO_DISABLE_MODELS_FETCH disables both the overlay and the fetch.
+func EnableModelsDevOverlay() {
+	modelsDevEnabled.Store(true)
+}
+
+// cachedModelsDevProviders loads the cached snapshot once per process. Not
+// enabled, missing, stale (> modelsDevMaxAge), or malformed all yield nil and
+// the curated catalog is used untouched. Read once deliberately:
+// DefaultRegistry is called on hot paths (pickers, cost views) and must not
+// re-stat the file every time; a background refresh benefits the NEXT process.
+func cachedModelsDevProviders() map[string]map[string]modelsDevModel {
+	if !modelsDevEnabled.Load() || strings.TrimSpace(os.Getenv("ZERO_DISABLE_MODELS_FETCH")) != "" {
+		return nil
+	}
+	modelsDevOnce.Do(func() {
+		path, err := modelsDevCachePath()
+		if err != nil {
+			return
+		}
+		info, err := os.Stat(path)
+		if err != nil || time.Since(info.ModTime()) > modelsDevMaxAge {
+			return
+		}
+		data, err := os.ReadFile(path)
+		if err != nil {
+			return
+		}
+		providers, err := parseModelsDev(data)
+		if err != nil {
+			return
+		}
+		modelsDevCached = providers
+	})
+	return modelsDevCached
+}
+
+// resetModelsDevCacheForTest clears the process-level cache memoization and
+// disables the overlay.
+func resetModelsDevCacheForTest() {
+	modelsDevOnce = sync.Once{}
+	modelsDevCached = nil
+	modelsDevEnabled.Store(false)
+}
+
+// RefreshModelsDevCache fetches models.dev/api.json into the on-disk cache
+// when the cache is missing or older than modelsDevRefreshAfter. It is safe to
+// call fire-and-forget from startup (use a goroutine); it never affects the
+// current process's registry (see cachedModelsDevProviders). Disabled entirely
+// by ZERO_DISABLE_MODELS_FETCH. The URL can be overridden with ZERO_MODELS_URL.
+func RefreshModelsDevCache(ctx context.Context) error {
+	if strings.TrimSpace(os.Getenv("ZERO_DISABLE_MODELS_FETCH")) != "" {
+		return nil
+	}
+	path, err := modelsDevCachePath()
+	if err != nil {
+		return err
+	}
+	if info, err := os.Stat(path); err == nil && time.Since(info.ModTime()) < modelsDevRefreshAfter {
+		return nil
+	}
+
+	url := strings.TrimSpace(os.Getenv("ZERO_MODELS_URL"))
+	if url == "" {
+		url = modelsDevDefaultURL
+	}
+	fetchCtx, cancel := context.WithTimeout(ctx, modelsDevFetchWindow)
+	defer cancel()
+	request, err := http.NewRequestWithContext(fetchCtx, http.MethodGet, url, nil)
+	if err != nil {
+		return err
+	}
+	request.Header.Set("User-Agent", "zero-models-refresh/0.1")
+	response, err := http.DefaultClient.Do(request)
+	if err != nil {
+		return err
+	}
+	defer response.Body.Close()
+	if response.StatusCode != http.StatusOK {
+		return fmt.Errorf("modelregistry: models.dev fetch: HTTP %d", response.StatusCode)
+	}
+	data, err := io.ReadAll(io.LimitReader(response.Body, modelsDevFetchLimit))
+	if err != nil {
+		return err
+	}
+	// Validate before persisting: a bad body must never clobber a good cache.
+	if _, err := parseModelsDev(data); err != nil {
+		return err
+	}
+	if err := os.MkdirAll(filepath.Dir(path), 0o755); err != nil {
+		return err
+	}
+	temp, err := os.CreateTemp(filepath.Dir(path), "modelsdev-*.json")
+	if err != nil {
+		return err
+	}
+	if _, err := temp.Write(data); err != nil {
+		_ = temp.Close()
+		_ = os.Remove(temp.Name())
+		return err
+	}
+	if err := temp.Close(); err != nil {
+		_ = os.Remove(temp.Name())
+		return err
+	}
+	return os.Rename(temp.Name(), path)
+}

internal/modelregistry/modelsdev_test.go

@@ -0,0 +1,228 @@
+package modelregistry
+
+import (
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"path/filepath"
+	"testing"
+	"time"
+)
+
+const sampleModelsDev = `{
+  "anthropic": {
+    "id": "anthropic",
+    "models": {
+      "claude-sonnet-4-5-20250929": {
+        "limit": {"context": 1000000, "output": 64000},
+        "cost": {"input": 3.5, "output": 17.5, "cache_read": 0.35, "cache_write": 4.4}
+      }
+    }
+  },
+  "google": {
+    "id": "google",
+    "models": {
+      "gemini-2.5-pro": {
+        "limit": {"context": 2097152, "output": 65536},
+        "cost": {"input": 9.99, "output": 9.99}
+      }
+    }
+  }
+}`
+
+func TestParseModelsDev(t *testing.T) {
+	providers, err := parseModelsDev([]byte(sampleModelsDev))
+	if err != nil {
+		t.Fatal(err)
+	}
+	record, ok := providers["anthropic"]["claude-sonnet-4-5-20250929"]
+	if !ok {
+		t.Fatal("expected anthropic sonnet record")
+	}
+	if record.Limit.Context != 1_000_000 || record.Cost.Input != 3.5 {
+		t.Fatalf("unexpected record: %+v", record)
+	}
+	if _, err := parseModelsDev([]byte(`{}`)); err == nil {
+		t.Fatal("empty document must be rejected")
+	}
+	if _, err := parseModelsDev([]byte(`not json`)); err == nil {
+		t.Fatal("malformed document must be rejected")
+	}
+}
+
+func TestApplyModelsDevOverrides(t *testing.T) {
+	// Point the cache at a non-existent file so DefaultModelEntries returns the
+	// pure curated catalog, then apply the sample snapshot explicitly.
+	t.Setenv("ZERO_MODELS_CACHE_PATH", filepath.Join(t.TempDir(), "absent.json"))
+	resetModelsDevCacheForTest()
+	t.Cleanup(resetModelsDevCacheForTest)
+
+	providers, err := parseModelsDev([]byte(sampleModelsDev))
+	if err != nil {
+		t.Fatal(err)
+	}
+	entries := applyModelsDevOverrides(DefaultModelEntries(), providers)
+
+	var sonnet, geminiPro, opus ModelEntry
+	for _, entry := range entries {
+		switch entry.ID {
+		case "claude-sonnet-4.5":
+			sonnet = entry
+		case "gemini-2.5-pro":
+			geminiPro = entry
+		case "claude-opus-4.1":
+			opus = entry
+		}
+	}
+
+	// Known model: limits and base pricing refreshed from the snapshot.
+	if sonnet.ContextLimits.ContextWindow != 1_000_000 || sonnet.ContextLimits.MaxOutputTokens != 64_000 {
+		t.Fatalf("sonnet limits not overridden: %+v", sonnet.ContextLimits)
+	}
+	if sonnet.Cost.InputPerMillion != 3.5 || sonnet.Cost.OutputPerMillion != 17.5 || sonnet.Cost.CacheWritePerMillion != 4.4 {
+		t.Fatalf("sonnet cost not overridden: %+v", sonnet.Cost)
+	}
+	if sonnet.Cost.Source != "models.dev/api.json (cached)" {
+		t.Fatalf("sonnet cost source not marked: %q", sonnet.Cost.Source)
+	}
+
+	// Tiered pricing is curated: limits refresh, cost must NOT (gemini-2.5-pro
+	// has curated tiers and the snapshot's flat 9.99 would misprice them).
+	if geminiPro.ContextLimits.ContextWindow != 2_097_152 {
+		t.Fatalf("gemini limits not overridden: %+v", geminiPro.ContextLimits)
+	}
+	if geminiPro.Cost.InputPerMillion == 9.99 || len(geminiPro.Cost.Tiers) == 0 {
+		t.Fatalf("tiered cost must stay curated: %+v", geminiPro.Cost)
+	}
+
+	// Model absent from the snapshot: untouched.
+	if opus.ContextLimits.ContextWindow != 200_000 || opus.Cost.InputPerMillion != 15 {
+		t.Fatalf("opus must be untouched: %+v %+v", opus.ContextLimits, opus.Cost)
+	}
+}
+
+func TestRefreshModelsDevCacheFetchesAndCaches(t *testing.T) {
+	hits := 0
+	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		hits++
+		_, _ = w.Write([]byte(sampleModelsDev))
+	}))
+	defer server.Close()
+
+	cachePath := filepath.Join(t.TempDir(), "modelsdev.json")
+	t.Setenv("ZERO_MODELS_CACHE_PATH", cachePath)
+	t.Setenv("ZERO_MODELS_URL", server.URL)
+	t.Setenv("ZERO_DISABLE_MODELS_FETCH", "")
+
+	if err := RefreshModelsDevCache(t.Context()); err != nil {
+		t.Fatal(err)
+	}
+	if hits != 1 {
+		t.Fatalf("expected 1 fetch, got %d", hits)
+	}
+	if _, err := os.Stat(cachePath); err != nil {
+		t.Fatalf("cache file missing: %v", err)
+	}
+	// Fresh cache: second call must not re-fetch.
+	if err := RefreshModelsDevCache(t.Context()); err != nil {
+		t.Fatal(err)
+	}
+	if hits != 1 {
+		t.Fatalf("fresh cache must skip the fetch, got %d hits", hits)
+	}
+}
+
+func TestRefreshModelsDevCacheRejectsBadBodyWithoutClobbering(t *testing.T) {
+	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		_, _ = w.Write([]byte("not json"))
+	}))
+	defer server.Close()
+
+	cachePath := filepath.Join(t.TempDir(), "modelsdev.json")
+	if err := os.WriteFile(cachePath, []byte(sampleModelsDev), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	stale := time.Now().Add(-48 * time.Hour)
+	if err := os.Chtimes(cachePath, stale, stale); err != nil {
+		t.Fatal(err)
+	}
+	t.Setenv("ZERO_MODELS_CACHE_PATH", cachePath)
+	t.Setenv("ZERO_MODELS_URL", server.URL)
+
+	if err := RefreshModelsDevCache(t.Context()); err == nil {
+		t.Fatal("bad body must return an error")
+	}
+	content, err := os.ReadFile(cachePath)
+	if err != nil || string(content) != sampleModelsDev {
+		t.Fatal("bad fetch must not clobber the existing cache")
+	}
+}
+
+func TestRefreshModelsDevCacheDisabledByEnv(t *testing.T) {
+	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		t.Fatal("fetch must not happen when disabled")
+	}))
+	defer server.Close()
+	t.Setenv("ZERO_MODELS_CACHE_PATH", filepath.Join(t.TempDir(), "modelsdev.json"))
+	t.Setenv("ZERO_MODELS_URL", server.URL)
+	t.Setenv("ZERO_DISABLE_MODELS_FETCH", "1")
+	if err := RefreshModelsDevCache(t.Context()); err != nil {
+		t.Fatal(err)
+	}
+}
+
+func TestCachedModelsDevProvidersIgnoresStaleCache(t *testing.T) {
+	cachePath := filepath.Join(t.TempDir(), "modelsdev.json")
+	if err := os.WriteFile(cachePath, []byte(sampleModelsDev), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	stale := time.Now().Add(-modelsDevMaxAge - time.Hour)
+	if err := os.Chtimes(cachePath, stale, stale); err != nil {
+		t.Fatal(err)
+	}
+	t.Setenv("ZERO_MODELS_CACHE_PATH", cachePath)
+	resetModelsDevCacheForTest()
+	t.Cleanup(resetModelsDevCacheForTest)
+	EnableModelsDevOverlay()
+
+	if providers := cachedModelsDevProviders(); providers != nil {
+		t.Fatal("stale cache must be ignored")
+	}
+}
+
+func TestCachedModelsDevProvidersRequiresOptIn(t *testing.T) {
+	cachePath := filepath.Join(t.TempDir(), "modelsdev.json")
+	if err := os.WriteFile(cachePath, []byte(sampleModelsDev), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	t.Setenv("ZERO_MODELS_CACHE_PATH", cachePath)
+	resetModelsDevCacheForTest()
+	t.Cleanup(resetModelsDevCacheForTest)
+
+	// Without EnableModelsDevOverlay a fresh, valid cache must still be ignored:
+	// library consumers and hermetic tests get the pure curated catalog.
+	if providers := cachedModelsDevProviders(); providers != nil {
+		t.Fatal("overlay must be opt-in")
+	}
+}
+
+func TestDefaultModelEntriesAppliesFreshCache(t *testing.T) {
+	cachePath := filepath.Join(t.TempDir(), "modelsdev.json")
+	if err := os.WriteFile(cachePath, []byte(sampleModelsDev), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	t.Setenv("ZERO_MODELS_CACHE_PATH", cachePath)
+	resetModelsDevCacheForTest()
+	t.Cleanup(resetModelsDevCacheForTest)
+	EnableModelsDevOverlay()
+
+	for _, entry := range DefaultModelEntries() {
+		if entry.ID == "claude-sonnet-4.5" {
+			if entry.ContextLimits.ContextWindow != 1_000_000 {
+				t.Fatalf("fresh cache must overlay the registry: %+v", entry.ContextLimits)
+			}
+			return
+		}
+	}
+	t.Fatal("claude-sonnet-4.5 not found")
+}

internal/providers/anthropic/cache_breakpoints_test.go

@@ -0,0 +1,96 @@
+package anthropic
+
+import (
+	"context"
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/Gitlawb/zero/internal/zeroruntime"
+)
+
+// The final two messages must carry a cache_control breakpoint on their last
+// cacheable block so the conversation transcript — not just system prompt and
+// tools — is a prompt-cache hit on the next turn. Earlier messages must stay
+// unmarked (Anthropic caps breakpoints at 4 per request: system, tools, and
+// these two).
+func TestAnthropicRequestMarksLastTwoMessagesForCaching(t *testing.T) {
+	var gotBody map[string]any
+	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		if err := json.NewDecoder(r.Body).Decode(&gotBody); err != nil {
+			t.Fatalf("decode request body: %v", err)
+		}
+		writeSSEEvent(w, "message_stop", `{"type":"message_stop"}`)
+	}))
+	defer server.Close()
+
+	provider, err := New(Options{BaseURL: server.URL + "/", Model: "claude-test"})
+	if err != nil {
+		t.Fatalf("New returned error: %v", err)
+	}
+	stream, err := provider.StreamCompletion(context.Background(), zeroruntime.CompletionRequest{
+		Messages: []zeroruntime.Message{
+			{Role: zeroruntime.MessageRoleUser, Content: "first"},
+			{
+				Role:      zeroruntime.MessageRoleAssistant,
+				Content:   "calling a tool",
+				ToolCalls: []zeroruntime.ToolCall{{ID: "toolu_1", Name: "grep", Arguments: `{}`}},
+			},
+			{Role: zeroruntime.MessageRoleTool, Content: "result", ToolCallID: "toolu_1"},
+		},
+	})
+	if err != nil {
+		t.Fatalf("StreamCompletion returned error: %v", err)
+	}
+	drain(stream)
+
+	messages := gotBody["messages"].([]any)
+	if len(messages) != 3 {
+		t.Fatalf("expected 3 wire messages, got %d: %#v", len(messages), messages)
+	}
+	lastBlockHasCache := func(message any) bool {
+		blocks, ok := message.(map[string]any)["content"].([]any)
+		if !ok || len(blocks) == 0 {
+			return false
+		}
+		last := blocks[len(blocks)-1].(map[string]any)
+		control, ok := last["cache_control"].(map[string]any)
+		return ok && control["type"] == "ephemeral"
+	}
+	if lastBlockHasCache(messages[0]) {
+		t.Fatalf("first message must not carry a breakpoint: %#v", messages[0])
+	}
+	if !lastBlockHasCache(messages[1]) {
+		t.Fatalf("second-to-last message must carry a breakpoint: %#v", messages[1])
+	}
+	if !lastBlockHasCache(messages[2]) {
+		t.Fatalf("last message must carry a breakpoint: %#v", messages[2])
+	}
+}
+
+// A string-content message must be converted to a block array so it can carry
+// the breakpoint, and thinking blocks must never carry cache_control (the API
+// rejects them) — the marker goes on the last cacheable block instead.
+func TestApplyMessageCacheBreakpointsSkipsThinkingBlocks(t *testing.T) {
+	messages := []anthropicMessage{
+		{Role: "user", Content: "plain string"},
+		{Role: "assistant", Content: []map[string]any{
+			{"type": "text", "text": "answer"},
+			{"type": "thinking", "thinking": "hmm", "signature": "sig"},
+		}},
+	}
+	applyMessageCacheBreakpoints(messages)
+
+	userBlocks := contentBlocks(messages[0].Content)
+	if len(userBlocks) != 1 || userBlocks[0]["cache_control"] == nil {
+		t.Fatalf("string content must become a marked block: %#v", messages[0].Content)
+	}
+	assistantBlocks := contentBlocks(messages[1].Content)
+	if assistantBlocks[1]["cache_control"] != nil {
+		t.Fatalf("thinking block must not be marked: %#v", assistantBlocks[1])
+	}
+	if assistantBlocks[0]["cache_control"] == nil {
+		t.Fatalf("text block must carry the breakpoint instead: %#v", assistantBlocks[0])
+	}
+}

internal/providers/anthropic/provider.go

@@ -405,9 +405,35 @@ func (provider *Provider) anthropicRequest(request zeroruntime.CompletionRequest
 		}
 		mapped.Tools[len(mapped.Tools)-1].CacheControl = &cacheControl{Type: cacheEphemeral}
 	}
+	applyMessageCacheBreakpoints(mapped.Messages)
 	return mapped, nil
 }
 
+// applyMessageCacheBreakpoints marks the last content block of the final two
+// messages with cache_control so the conversation transcript is cached
+// turn-over-turn, not just the system prompt and tool definitions. Two message
+// breakpoints (not one) keep the previous turn's prefix a cache hit while the
+// newest suffix is being written. Anthropic allows at most 4 breakpoints per
+// request: system + tools use two above, these use the remaining two. Thinking
+// blocks cannot carry cache_control, so the marker goes on the last block that
+// can.
+func applyMessageCacheBreakpoints(messages []anthropicMessage) {
+	marked := 0
+	for i := len(messages) - 1; i >= 0 && marked < 2; i-- {
+		blocks := contentBlocks(messages[i].Content)
+		for j := len(blocks) - 1; j >= 0; j-- {
+			blockType, _ := blocks[j]["type"].(string)
+			if blockType == "thinking" || blockType == "redacted_thinking" {
+				continue
+			}
+			blocks[j]["cache_control"] = map[string]any{"type": cacheEphemeral}
+			messages[i].Content = blocks
+			marked++
+			break
+		}
+	}
+}
+
 func mapMessages(messages []zeroruntime.Message) (string, []anthropicMessage, error) {
 	systemParts := []string{}
 	mapped := []anthropicMessage{}

internal/providers/openai/provider.go

@@ -9,6 +9,7 @@ import (
 	"io"
 	"net/http"
 	"net/url"
+	"os"
 	"strings"
 	"time"
 
@@ -283,10 +284,10 @@ func (provider *Provider) emitChunk(
 	events chan<- zeroruntime.StreamEvent,
 ) {
 	for _, choice := range chunk.Choices {
-		if choice.Delta.ReasoningContent != "" {
+		if reasoning := choice.Delta.reasoningText(); reasoning != "" {
 			sendEvent(ctx, events, zeroruntime.StreamEvent{
 				Type:    zeroruntime.StreamEventReasoning,
-				Content: choice.Delta.ReasoningContent,
+				Content: reasoning,
 			})
 		}
 		if choice.Delta.Content != "" {
@@ -317,6 +318,13 @@ func (provider *Provider) emitChunk(
 	}
 }
 
+func (delta streamDelta) reasoningText() string {
+	if delta.ReasoningContent != "" {
+		return delta.ReasoningContent
+	}
+	return delta.Reasoning
+}
+
 // mapFinishReason maps OpenAI's finish_reason onto the runtime's normalized
 // terminal reasons. A normal finish ("stop"/"tool_calls"/"") returns "".
 func mapFinishReason(reason string) string {
@@ -414,6 +422,12 @@ func (provider *Provider) openAIRequest(request zeroruntime.CompletionRequest) c
 	if effort := openAIReasoningEffort(request.ReasoningEffort); effort != "" {
 		mapped.ReasoningEffort = effort
 	}
+	// prompt_cache_key is a documented OpenAI parameter; compatible servers
+	// ignore unknown fields, but a strict endpoint that rejects it can be
+	// accommodated with ZERO_DISABLE_PROMPT_CACHE_KEY=1.
+	if key := strings.TrimSpace(request.PromptCacheKey); key != "" && !promptCacheKeyDisabled() {
+		mapped.PromptCacheKey = key
+	}
 	if len(request.Tools) > 0 {
 		mapped.Tools = make([]toolDefinition, 0, len(request.Tools))
 		for _, tool := range request.Tools {
@@ -430,6 +444,15 @@ func (provider *Provider) openAIRequest(request zeroruntime.CompletionRequest) c
 	return mapped
 }
 
+// promptCacheKeyDisabled reports whether the ZERO_DISABLE_PROMPT_CACHE_KEY
+// kill switch is set to a truthy value. "0" and "false" (any case) are
+// no-ops, matching how ZERO_FORMAT_ON_WRITE parses boolean flags, so an
+// explicitly-disabled toggle never flips the behavior it names.
+func promptCacheKeyDisabled() bool {
+	value := strings.TrimSpace(os.Getenv("ZERO_DISABLE_PROMPT_CACHE_KEY"))
+	return value != "" && value != "0" && !strings.EqualFold(value, "false")
+}
+
 // openAIReasoningEffort normalizes a requested effort to a value the OpenAI chat
 // completions API accepts, or "" to omit the field. "none" (and anything else)
 // is dropped rather than risking a 400 on an unrecognized enum.

internal/providers/openai/provider_test.go

@@ -227,6 +227,39 @@ func TestStreamCompletionEmitsReasoningContentDeltas(t *testing.T) {
 	}
 }
 
+func TestStreamCompletionEmitsReasoningAliasDeltas(t *testing.T) {
+	provider := newTestProvider(t, func(w http.ResponseWriter, r *http.Request) {
+		writeSSE(w, `{"choices":[{"delta":{"reasoning":"Thinking. "}}]}`)
+		writeSSE(w, `{"choices":[{"delta":{"reasoning":"Answering now."}}]}`)
+		writeSSE(w, `[DONE]`)
+	})
+
+	events := collectProviderEvents(t, provider)
+	reasoning := eventsOfType(events, zeroruntime.StreamEventReasoning)
+	if len(reasoning) != 2 {
+		t.Fatalf("reasoning events = %#v, want two reasoning deltas", reasoning)
+	}
+	if reasoning[0].Content != "Thinking. " || reasoning[1].Content != "Answering now." {
+		t.Fatalf("unexpected reasoning events: %#v", reasoning)
+	}
+	if text := eventsOfType(events, zeroruntime.StreamEventText); len(text) != 0 {
+		t.Fatalf("reasoning must not emit text events, got %#v", text)
+	}
+}
+
+func TestStreamCompletionPrefersReasoningContentOverAlias(t *testing.T) {
+	provider := newTestProvider(t, func(w http.ResponseWriter, r *http.Request) {
+		writeSSE(w, `{"choices":[{"delta":{"reasoning_content":"standard","reasoning":"alias"}}]}`)
+		writeSSE(w, `[DONE]`)
+	})
+
+	events := collectProviderEvents(t, provider)
+	reasoning := eventsOfType(events, zeroruntime.StreamEventReasoning)
+	if len(reasoning) != 1 || reasoning[0].Content != "standard" {
+		t.Fatalf("reasoning events = %#v, want standard reasoning_content", reasoning)
+	}
+}
+
 func TestStreamCompletionEmitsReasoningBeforeRegularContent(t *testing.T) {
 	provider := newTestProvider(t, func(w http.ResponseWriter, r *http.Request) {
 		writeSSE(w, `{"choices":[{"delta":{"reasoning_content":"Thinking. ","content":"Answer."}}]}`)
@@ -1226,3 +1259,61 @@ func TestOpenAIRequestEmptyContentHandling(t *testing.T) {
 		t.Fatalf("assistant tool calls dropped: %#v", req.Messages[1])
 	}
 }
+
+// TestOpenAIRequestPromptCacheKey locks in prompt_cache_key forwarding: a
+// session-carrying request serializes the key so the backend can route to a
+// replica holding the cached prefix, a keyless request omits the field
+// entirely (strict servers see byte-identical requests to before), and
+// ZERO_DISABLE_PROMPT_CACHE_KEY suppresses it for endpoints that reject it.
+func TestOpenAIRequestPromptCacheKey(t *testing.T) {
+	provider, err := New(Options{Model: "gpt-test"})
+	if err != nil {
+		t.Fatalf("New returned error: %v", err)
+	}
+	messages := []zeroruntime.Message{{Role: zeroruntime.MessageRoleUser, Content: "hi"}}
+
+	req := provider.openAIRequest(zeroruntime.CompletionRequest{
+		Messages:       messages,
+		PromptCacheKey: "sess_123",
+	})
+	if req.PromptCacheKey != "sess_123" {
+		t.Fatalf("PromptCacheKey = %q, want sess_123", req.PromptCacheKey)
+	}
+	data, err := json.Marshal(req)
+	if err != nil {
+		t.Fatalf("marshal: %v", err)
+	}
+	if !strings.Contains(string(data), `"prompt_cache_key":"sess_123"`) {
+		t.Fatalf("prompt_cache_key not serialized: %s", data)
+	}
+
+	req = provider.openAIRequest(zeroruntime.CompletionRequest{Messages: messages})
+	if data, err = json.Marshal(req); err != nil {
+		t.Fatalf("marshal: %v", err)
+	}
+	if strings.Contains(string(data), "prompt_cache_key") {
+		t.Fatalf("keyless request must omit prompt_cache_key: %s", data)
+	}
+
+	t.Setenv("ZERO_DISABLE_PROMPT_CACHE_KEY", "1")
+	req = provider.openAIRequest(zeroruntime.CompletionRequest{
+		Messages:       messages,
+		PromptCacheKey: "sess_123",
+	})
+	if req.PromptCacheKey != "" {
+		t.Fatalf("kill switch ignored; PromptCacheKey = %q", req.PromptCacheKey)
+	}
+
+	// Explicitly-falsy kill switch values must NOT disable forwarding — only
+	// truthy values flip the toggle (same parsing as ZERO_FORMAT_ON_WRITE).
+	for _, value := range []string{"0", "false", "FALSE"} {
+		t.Setenv("ZERO_DISABLE_PROMPT_CACHE_KEY", value)
+		req = provider.openAIRequest(zeroruntime.CompletionRequest{
+			Messages:       messages,
+			PromptCacheKey: "sess_123",
+		})
+		if req.PromptCacheKey != "sess_123" {
+			t.Fatalf("ZERO_DISABLE_PROMPT_CACHE_KEY=%q must be a no-op; PromptCacheKey = %q", value, req.PromptCacheKey)
+		}
+	}
+}

internal/providers/openai/types.go

@@ -8,6 +8,11 @@ type chatCompletionRequest struct {
 	ReasoningEffort     string           `json:"reasoning_effort,omitempty"`
 	Stream              bool             `json:"stream"`
 	StreamOptions       *streamOptions   `json:"stream_options,omitempty"`
+	// PromptCacheKey asks the backend to route the request to a replica that
+	// already holds this conversation's prefix in its prompt cache (the OpenAI
+	// `prompt_cache_key` parameter). Omitted when the caller carries no session
+	// identity or when ZERO_DISABLE_PROMPT_CACHE_KEY is set.
+	PromptCacheKey string `json:"prompt_cache_key,omitempty"`
 }
 
 // streamOptions requests the final usage chunk on a streaming response. Without
@@ -78,6 +83,7 @@ type streamChoice struct {
 type streamDelta struct {
 	Content          string                `json:"content"`
 	ReasoningContent string                `json:"reasoning_content"`
+	Reasoning        string                `json:"reasoning"`
 	ToolCalls        []streamToolCallDelta `json:"tool_calls"`
 }
 

internal/providers/providerio/retry.go

@@ -23,12 +23,18 @@ import (
 // (billable) work. Only the INITIAL request is ever in scope; once the response
 // body starts streaming it is never re-issued.
 
-const defaultMaxRetryAttempts = 3
+const defaultMaxRetryAttempts = 6
 
 // maxBackoff caps a single backoff wait so a hostile or buggy Retry-After can't
 // stall the agent for minutes.
 const maxBackoff = 30 * time.Second
 
+// retryBackoffBase is the first wait when the server supplied no Retry-After.
+// Rate-limit windows are measured in seconds, not milliseconds: retrying a 429
+// after 400ms almost always burns the attempt while still limited, so the
+// schedule is 2s, 4s, 8s, 16s, then maxBackoff. A var so tests can shrink it.
+var retryBackoffBase = 2 * time.Second
+
 // SendWithRetry issues an HTTP request, retrying ONLY the safe-to-replay server
 // responses (429 and 503, see ShouldRetryStatus) up to maxAttempts — backing off
 // between tries and honoring a server Retry-After header and context
@@ -110,17 +116,11 @@ func ShouldRetryStatus(code int) bool {
 }
 
 // Backoff waits before retry attempt N (1-based), returning false if the context
-// is cancelled during the wait. The wait is attempt*400ms unless the server
-// supplied a (positive) Retry-After, and is capped at maxBackoff.
+// is cancelled during the wait. A server-supplied (positive) Retry-After wins;
+// otherwise the wait doubles from retryBackoffBase per attempt. Either way the
+// wait is capped at maxBackoff.
 func Backoff(ctx context.Context, attempt int, retryAfter time.Duration) bool {
-	wait := time.Duration(attempt) * 400 * time.Millisecond
-	if retryAfter > 0 {
-		wait = retryAfter
-	}
-	if wait > maxBackoff {
-		wait = maxBackoff
-	}
-	timer := time.NewTimer(wait)
+	timer := time.NewTimer(backoffWait(attempt, retryAfter))
 	defer timer.Stop()
 	select {
 	case <-ctx.Done():
@@ -130,6 +130,27 @@ func Backoff(ctx context.Context, attempt int, retryAfter time.Duration) bool {
 	}
 }
 
+// backoffWait computes the wait before retry attempt N (1-based): Retry-After
+// when supplied, else exponential from retryBackoffBase, both capped at
+// maxBackoff. The exponent is clamped so a large attempt count cannot overflow.
+func backoffWait(attempt int, retryAfter time.Duration) time.Duration {
+	wait := retryAfter
+	if wait <= 0 {
+		exponent := attempt - 1
+		if exponent > 5 {
+			exponent = 5
+		}
+		if exponent < 0 {
+			exponent = 0
+		}
+		wait = retryBackoffBase * time.Duration(1<<exponent)
+	}
+	if wait > maxBackoff {
+		wait = maxBackoff
+	}
+	return wait
+}
+
 // RetryAfter parses a response's Retry-After header (delay-seconds or an HTTP
 // date) into a positive duration, or 0 when absent/unparseable. The result is
 // capped at maxBackoff by Backoff.

internal/providers/providerio/retry_test.go

@@ -94,7 +94,40 @@ func TestBackoffWaitsThenReturnsTrue(t *testing.T) {
 	}
 }
 
+// shrinkBackoff makes retry waits negligible for the duration of a test.
+func shrinkBackoff(t *testing.T) {
+	t.Helper()
+	saved := retryBackoffBase
+	retryBackoffBase = time.Millisecond
+	t.Cleanup(func() { retryBackoffBase = saved })
+}
+
+func TestBackoffWaitSchedule(t *testing.T) {
+	// Without Retry-After the wait doubles per attempt from 2s and caps at 30s;
+	// a supplied Retry-After wins but is capped too.
+	cases := []struct {
+		attempt    int
+		retryAfter time.Duration
+		want       time.Duration
+	}{
+		{1, 0, 2 * time.Second},
+		{2, 0, 4 * time.Second},
+		{3, 0, 8 * time.Second},
+		{4, 0, 16 * time.Second},
+		{5, 0, 30 * time.Second},  // 32s capped
+		{50, 0, 30 * time.Second}, // clamped exponent, no overflow
+		{1, 7 * time.Second, 7 * time.Second},
+		{1, 5 * time.Minute, 30 * time.Second}, // hostile Retry-After capped
+	}
+	for _, c := range cases {
+		if got := backoffWait(c.attempt, c.retryAfter); got != c.want {
+			t.Errorf("backoffWait(%d, %v) = %v, want %v", c.attempt, c.retryAfter, got, c.want)
+		}
+	}
+}
+
 func TestSendWithRetryRetriesThenSucceeds(t *testing.T) {
+	shrinkBackoff(t)
 	var hits int32
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		if atomic.AddInt32(&hits, 1) == 1 {
@@ -140,6 +173,7 @@ func TestSendWithRetryReturnsNonRetryableImmediately(t *testing.T) {
 }
 
 func TestSendWithRetryReturnsLastResponseAfterMaxAttempts(t *testing.T) {
+	shrinkBackoff(t)
 	var hits int32
 	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		atomic.AddInt32(&hits, 1)

internal/repoinfo/repoinfo.go

@@ -54,6 +54,9 @@ type Info struct {
 	AgeDays               *int       `json:"ageDays,omitempty"`
 	Contributors90d       *int       `json:"contributors90d,omitempty"`
 	CommitVelocity30d     *int       `json:"commitVelocity30d,omitempty"`
+	BranchCount           *int       `json:"branchCount,omitempty"`
+	CommitCount           *int       `json:"commitCount,omitempty"`
+	TagCount              *int       `json:"tagCount,omitempty"`
 }
 
 // RunGit runs a git subcommand in dir and returns raw stdout. A non-zero exit
@@ -229,6 +232,43 @@ func Collect(ctx context.Context, opts Options) (Info, error) {
 			info.CommitVelocity30d = &n
 		}
 	}
+	if count, err := run(ctx, dir, "rev-list", "--count", "HEAD"); err == nil {
+		if n, perr := strconv.Atoi(strings.TrimSpace(count)); perr == nil {
+			info.CommitCount = &n
+		}
+	}
+	if branches, err := run(ctx, dir, "branch", "-a"); err == nil {
+		set := map[string]bool{}
+		for _, line := range strings.Split(branches, "\n") {
+			line = strings.TrimSpace(line)
+			if line == "" || strings.Contains(line, "->") {
+				continue
+			}
+			// Strip leading '* ' (current branch) or '+ ' (checked out in another worktree).
+			line = strings.TrimPrefix(line, "* ")
+			line = strings.TrimPrefix(line, "+ ")
+			line = strings.TrimSpace(line)
+			if line == "" || strings.HasPrefix(line, "(") {
+				continue
+			}
+			// If it's a remote branch, extract the branch name.
+			// E.g., remotes/origin/main -> main
+			// E.g., remotes/origin/feature/xyz -> feature/xyz
+			if strings.HasPrefix(line, "remotes/") {
+				line = strings.TrimPrefix(line, "remotes/")
+				if slashIdx := strings.Index(line, "/"); slashIdx != -1 {
+					line = line[slashIdx+1:]
+				}
+			}
+			set[line] = true
+		}
+		n := len(set)
+		info.BranchCount = &n
+	}
+	if tags, err := run(ctx, dir, "tag"); err == nil {
+		n := countNonEmptyLines(tags)
+		info.TagCount = &n
+	}
 
 	return info, nil
 }
@@ -328,3 +368,13 @@ func defaultRunGit(ctx context.Context, dir string, args ...string) (string, err
 	}
 	return stdout.String(), nil
 }
+
+func countNonEmptyLines(s string) int {
+	count := 0
+	for _, line := range strings.Split(s, "\n") {
+		if strings.TrimSpace(line) != "" {
+			count++
+		}
+	}
+	return count
+}

internal/repoinfo/repoinfo_test.go

@@ -43,6 +43,8 @@ func TestCollectCoreMetrics(t *testing.T) {
 			"remote":    "https://github.com/x/y.git\n",
 			"log":       "0\n",
 			"rev-list":  "12\n",
+			"branch":    "  main\n  remotes/origin/HEAD -> origin/main\n  remotes/origin/main\n  remotes/origin/feature\n",
+			"tag":       "v0.1.0\nv0.2.0\n",
 		}, nil),
 	})
 	if err != nil {
@@ -85,6 +87,15 @@ func TestCollectCoreMetrics(t *testing.T) {
 	if info.CommitVelocity30d == nil || *info.CommitVelocity30d != 12 {
 		t.Fatalf("CommitVelocity30d=%v want 12", info.CommitVelocity30d)
 	}
+	if info.CommitCount == nil || *info.CommitCount != 12 {
+		t.Fatalf("CommitCount=%v want 12", info.CommitCount)
+	}
+	if info.BranchCount == nil || *info.BranchCount != 2 {
+		t.Fatalf("BranchCount=%v want 2", info.BranchCount)
+	}
+	if info.TagCount == nil || *info.TagCount != 2 {
+		t.Fatalf("TagCount=%v want 2", info.TagCount)
+	}
 }
 
 func TestCollectContributorsUnique(t *testing.T) {
@@ -252,12 +263,13 @@ func TestCollectNotGitRepo(t *testing.T) {
 }
 
 func TestCollectOnlyReadOnlyLocalSubcommands(t *testing.T) {
-	allowed := map[string]bool{"ls-tree": true, "rev-parse": true, "remote": true, "log": true, "rev-list": true}
+	allowed := map[string]bool{"ls-tree": true, "rev-parse": true, "remote": true, "log": true, "rev-list": true, "branch": true, "tag": true}
 	var used []string
 	_, err := Collect(context.Background(), Options{
 		Now: time.Unix(0, 0),
 		RunGit: fakeGit(t, map[string]string{
 			"ls-tree": lsTree, "rev-parse": "main\n", "remote": "u\n", "log": "0\n", "rev-list": "1\n",
+			"branch": "main\n", "tag": "v0.1.0\n",
 		}, &used),
 	})
 	if err != nil {

internal/sandbox/manager.go

@@ -3,8 +3,9 @@ package sandbox
 import (
 	"errors"
 	"os"
-	"os/exec"
+	"path/filepath"
 	"runtime"
+	"strings"
 )
 
 // errWindowsSandboxNotInitialized is returned only to a caller that explicitly
@@ -121,9 +122,55 @@ func (manager SandboxManager) Backend() Backend {
 	return manager.backend
 }
 
+// isExecutable checks whether a file is executable. On Unix, this checks the
+// execute permission bits. On Windows, Go's os.FileMode does not set execute
+// bits, so we check for common executable extensions instead.
+func isExecutable(fi os.FileInfo) bool {
+	if !fi.Mode().IsRegular() {
+		return false
+	}
+	if runtime.GOOS == "windows" {
+		name := strings.ToLower(fi.Name())
+		return strings.HasSuffix(name, ".exe") ||
+			strings.HasSuffix(name, ".com") ||
+			strings.HasSuffix(name, ".bat") ||
+			strings.HasSuffix(name, ".cmd")
+	}
+	return fi.Mode()&0o111 != 0
+}
+
+// lookupExecutable checks whether a named binary exists and is executable,
+// using os.Stat instead of exec.LookPath. exec.LookPath internally uses the
+// faccessat2 syscall, which is blocked by Android's seccomp filter (SIGSYS).
+func lookupExecutable(name string) (string, error) {
+	if !strings.Contains(name, string(os.PathSeparator)) {
+		path := os.Getenv("PATH")
+		for _, dir := range strings.Split(path, string(os.PathListSeparator)) {
+			if dir == "" {
+				continue // skip empty entries (e.g. trailing colon)
+			}
+			candidate := filepath.Join(dir, name)
+			if fi, err := os.Stat(candidate); err == nil {
+				if isExecutable(fi) {
+					return candidate, nil
+				}
+			}
+		}
+		return "", errors.New("executable file not found in $PATH")
+	}
+	fi, err := os.Stat(name)
+	if err != nil {
+		return "", err
+	}
+	if isExecutable(fi) {
+		return name, nil
+	}
+	return "", errors.New("executable file not found")
+}
+
 func selectPlatformBackend(goos string, lookup func(string) (string, error)) Backend {
 	if lookup == nil {
-		lookup = exec.LookPath
+		lookup = lookupExecutable
 	}
 	switch goos {
 	case "linux":

internal/sandbox/runner_windows_integration_test.go

@@ -190,6 +190,67 @@ func TestWindowsUnelevatedRealSandboxSmoke(t *testing.T) {
 	}
 }
 
+// TestWindowsRestrictedTokenNestedPipeCapture pins the fix in
+// broadenWindowsRestrictedTokenDefaultDacl (windows_token_windows.go): a
+// nested subprocess spawned FROM WITHIN the sandboxed process, with its
+// output captured via a pipe, must succeed. Before the fix, the sandboxed
+// process's own attempt to create a pipe for such a subprocess failed with
+// ERROR_ACCESS_DENIED (Win32 error 5) — the WRITE_RESTRICTED token's extra
+// access check has no restricted-SID match against a freshly created pipe's
+// default security descriptor. This is exactly the pattern any tool that
+// shells out internally and captures output hits (`gh` invoking `git`, for
+// one; cmd.exe's own FOR /F does the identical CreatePipe+CreateProcess
+// dance internally, so it reproduces the bug with no external dependency
+// beyond cmd.exe itself).
+func TestWindowsRestrictedTokenNestedPipeCapture(t *testing.T) {
+	if os.Getenv("ZERO_SANDBOX_REAL_SMOKE") != "1" {
+		t.Skip("set ZERO_SANDBOX_REAL_SMOKE=1 to run real Windows sandbox smoke tests")
+	}
+	runnerExe := realSmokeExecutable(t, "ZERO_WINDOWS_COMMAND_RUNNER_EXE", WindowsSandboxCommandRunnerName)
+
+	root := t.TempDir()
+	sandboxHome := filepath.Join(root, ".zero-sandbox")
+	profile := PermissionProfile{
+		FileSystem: FileSystemPolicy{
+			Kind:                 FileSystemRestricted,
+			ReadRoots:            []string{root},
+			WriteRoots:           []WritableRoot{{Root: root, ProtectedMetadataNames: []string{".git", ".zero", ".agents"}}},
+			IncludePlatformRoots: true,
+			AllowTemp:            true,
+		},
+		Network: NetworkPolicy{Mode: NetworkDeny},
+	}
+	config := WindowsSandboxCommandArgsOptions{
+		SandboxHome:       sandboxHome,
+		CommandCWD:        root,
+		WorkspaceRoots:    []string{root},
+		PermissionProfile: profile,
+		SandboxLevel:      WindowsSandboxLevelUnelevated,
+	}
+
+	marker := filepath.Join(root, "nested-pipe-marker.txt")
+	script := filepath.Join(root, "nested-pipe.cmd")
+	// FOR /F drives cmd.exe's own internal CreatePipe+CreateProcess capture
+	// of the quoted command's output — the exact mechanism this fix targets.
+	// The full path to whoami.exe sidesteps PATH resolution inside the
+	// sandboxed process's minimal environment.
+	scriptBody := "for /F %%i in ('C:\\Windows\\System32\\whoami.exe') do echo %%i> " + marker + "\r\n"
+	if err := os.WriteFile(script, []byte(scriptBody), 0o644); err != nil {
+		t.Fatalf("write nested-pipe script: %v", err)
+	}
+
+	runWindowsRealSmokeCommand(t, runnerExe, config, []string{
+		"cmd.exe", "/d", "/s", "/c", script,
+	}, 0)
+	captured, err := os.ReadFile(marker)
+	if err != nil {
+		t.Fatalf("read nested-pipe marker: %v", err)
+	}
+	if strings.TrimSpace(string(captured)) == "" {
+		t.Fatalf("nested-pipe marker is empty; FOR /F failed to capture the subprocess's output")
+	}
+}
+
 func realSmokeExecutable(t *testing.T, envKey string, fallbackName string) string {
 	t.Helper()
 	if path := os.Getenv(envKey); path != "" {

internal/sandbox/windows_token_windows.go

@@ -21,6 +21,11 @@ const (
 
 var procCreateRestrictedToken = windows.NewLazySystemDLL("advapi32.dll").NewProc("CreateRestrictedToken")
 
+// procSetEntriesInAclW merges new ACEs into an existing ACL. Not wrapped by
+// x/sys/windows (its internal setEntriesInAcl is unexported), so it is
+// declared directly the same way procCreateRestrictedToken is.
+var procSetEntriesInAclW = windows.NewLazySystemDLL("advapi32.dll").NewProc("SetEntriesInAclW")
+
 type windowsLocalSID struct {
 	sid *windows.SID
 }
@@ -122,9 +127,139 @@ func createWindowsRestrictedTokenFromBase(base windows.Token, capabilitySIDs []w
 		_ = restricted.Close()
 		return 0, err
 	}
+	if err := broadenWindowsRestrictedTokenDefaultDacl(restricted, sidFromBytes(logonSID)); err != nil {
+		_ = restricted.Close()
+		return 0, err
+	}
+	runtime.KeepAlive(logonSID)
 	return restricted, nil
 }
 
+// broadenWindowsRestrictedTokenDefaultDacl adds the token's own logon SID to
+// its default DACL (the DACL applied to any kernel object the token's process
+// creates WITHOUT an explicit security descriptor, e.g. anonymous pipes,
+// events, mutexes).
+//
+// A WRITE_RESTRICTED token requires a WRITE-type access check to match BOTH
+// the normal enabled-SID grant AND a SEPARATE grant to one of the token's
+// restricted SIDs. CreateRestrictedToken does not touch the default DACL it
+// inherits from the base token, so newly created objects only carry ACEs for
+// the base token's normal identity (owner, SYSTEM, Administrators) — none of
+// which are in the restricted-SID list — so the second check fails. This
+// breaks anything that creates a pipe/handle for itself with default
+// security, which is exactly what any tool using a language runtime's
+// "spawn a subprocess and capture its output" primitive does (Go's os/exec,
+// and by extension `gh`, and many other CLI tools that shell out
+// internally): CreatePipe() succeeds, but the creating process's own token
+// then fails to actually read/write the handle it just made.
+//
+// The logon SID is already unconditionally present in this token's
+// restricted-SID list (see createWindowsRestrictedTokenFromBase), so adding
+// it here — and ONLY it — to the default DACL is the minimal change that
+// satisfies the WRITE_RESTRICTED extra check for self-created objects. It
+// does not touch file-system access: NTFS write-jailing is enforced by the
+// explicit ACL grants applied to workspace paths (windows_acl.go), which are
+// created WITH an explicit security descriptor and never consult the token's
+// default DACL. The exposure is bounded to the same logon session (other
+// processes running as the same signed-in user could, in principle, open a
+// NAMED kernel object the sandboxed process creates with default security;
+// anonymous pipes — the actual object type this fixes — have no name and are
+// only reachable via an inherited handle, so this is a no-op for them).
+func broadenWindowsRestrictedTokenDefaultDacl(token windows.Token, logonSID *windows.SID) error {
+	oldDacl, oldDaclBuf, err := windowsTokenDefaultDacl(token)
+	if err != nil {
+		return fmt.Errorf("read token default DACL: %w", err)
+	}
+	// The second (WRITE-type) access check only needs a read/write match; a
+	// full GENERIC_ALL grant (which also implies DELETE, WRITE_DAC, and
+	// WRITE_OWNER) is broader than the pipe/event/mutex use case this exists
+	// for actually requires.
+	access := []windows.EXPLICIT_ACCESS{{
+		AccessPermissions: windows.GENERIC_READ | windows.GENERIC_WRITE,
+		AccessMode:        windows.GRANT_ACCESS,
+		Inheritance:       windows.NO_INHERITANCE,
+		Trustee: windows.TRUSTEE{
+			MultipleTrusteeOperation: windows.NO_MULTIPLE_TRUSTEE,
+			TrusteeForm:              windows.TRUSTEE_IS_SID,
+			TrusteeType:              windows.TRUSTEE_IS_USER,
+			TrusteeValue:             windows.TrusteeValueFromSID(logonSID),
+		},
+	}}
+	newDacl, err := setEntriesInACL(access, oldDacl)
+	// oldDacl points INTO oldDaclBuf (see windowsTokenDefaultDacl) and is
+	// dereferenced by native code inside setEntriesInACL, above. Nothing in Go
+	// references oldDaclBuf once windowsTokenDefaultDacl returned, so without
+	// this KeepAlive it is eligible for GC the moment an allocation in this
+	// function (the EXPLICIT_ACCESS literal, TrusteeValueFromSID) hits a
+	// safepoint, before the native call ever reads it - a real, if narrow,
+	// use-after-free window in security-boundary code.
+	runtime.KeepAlive(oldDaclBuf)
+	if err != nil {
+		return fmt.Errorf("merge token default DACL: %w", err)
+	}
+	defer func() { _, _ = windows.LocalFree(windows.Handle(unsafe.Pointer(newDacl))) }()
+	if err := windowsSetTokenDefaultDacl(token, newDacl); err != nil {
+		return fmt.Errorf("set token default DACL: %w", err)
+	}
+	return nil
+}
+
+type windowsTokenDefaultDaclInfo struct {
+	DefaultDacl *windows.ACL
+}
+
+// windowsTokenDefaultDacl returns the token's current default DACL along with
+// the buffer backing it. TOKEN_DEFAULT_DACL embeds the ACL data inside the
+// same allocation GetTokenInformation fills, so the returned *windows.ACL is
+// only valid for as long as the returned buffer stays reachable: the caller
+// must runtime.KeepAlive(buf) until every native call that dereferences the
+// ACL has completed. A KeepAlive inside this function, ending at its own
+// return, would not protect any use by the caller after that.
+func windowsTokenDefaultDacl(token windows.Token) (*windows.ACL, []byte, error) {
+	var size uint32
+	err := windows.GetTokenInformation(token, windows.TokenDefaultDacl, nil, 0, &size)
+	if err == nil || err != windows.ERROR_INSUFFICIENT_BUFFER {
+		return nil, nil, fmt.Errorf("size TokenDefaultDacl: %w", err)
+	}
+	buf := make([]byte, size)
+	if err := windows.GetTokenInformation(token, windows.TokenDefaultDacl, &buf[0], size, &size); err != nil {
+		return nil, nil, fmt.Errorf("get TokenDefaultDacl: %w", err)
+	}
+	info := (*windowsTokenDefaultDaclInfo)(unsafe.Pointer(&buf[0]))
+	return info.DefaultDacl, buf, nil
+}
+
+func windowsSetTokenDefaultDacl(token windows.Token, dacl *windows.ACL) error {
+	info := windowsTokenDefaultDaclInfo{DefaultDacl: dacl}
+	err := windows.SetTokenInformation(token, windows.TokenDefaultDacl, (*byte)(unsafe.Pointer(&info)), uint32(unsafe.Sizeof(info)))
+	runtime.KeepAlive(info)
+	return err
+}
+
+// setEntriesInACL merges entries into oldACL via Win32 SetEntriesInAclW,
+// returning a newly allocated ACL the caller must free with windows.LocalFree.
+func setEntriesInACL(entries []windows.EXPLICIT_ACCESS, oldACL *windows.ACL) (*windows.ACL, error) {
+	if len(entries) == 0 {
+		return nil, errors.New("setEntriesInACL requires at least one entry")
+	}
+	var newACL *windows.ACL
+	// SetEntriesInAclW reports failure directly in its return value (ERROR_SUCCESS
+	// on success, a Win32 error code otherwise); unlike many Win32 APIs it does not
+	// set the thread's last-error, so the syscall trio's second/third return
+	// values (which surface GetLastError) are not meaningful here.
+	ret, _, _ := procSetEntriesInAclW.Call(
+		uintptr(len(entries)),
+		uintptr(unsafe.Pointer(&entries[0])),
+		uintptr(unsafe.Pointer(oldACL)),
+		uintptr(unsafe.Pointer(&newACL)),
+	)
+	runtime.KeepAlive(entries)
+	if ret != 0 {
+		return nil, fmt.Errorf("SetEntriesInAclW: %w", syscall.Errno(ret))
+	}
+	return newACL, nil
+}
+
 func copyWindowsLogonSID(token windows.Token) ([]byte, error) {
 	groups, err := token.GetTokenGroups()
 	if err == nil {

internal/sessions/append_events_test.go

@@ -0,0 +1,233 @@
+package sessions
+
+import (
+	"encoding/json"
+	"fmt"
+	"reflect"
+	"strings"
+	"sync"
+	"testing"
+	"time"
+)
+
+func TestStoreAppendEventsBatchesSequencesAndMetadata(t *testing.T) {
+	store := NewStore(StoreOptions{RootDir: t.TempDir(), Now: sequenceClock([]time.Time{
+		time.Date(2026, 6, 4, 15, 0, 0, 0, time.UTC),
+		time.Date(2026, 6, 4, 15, 0, 1, 0, time.UTC),
+		time.Date(2026, 6, 4, 15, 0, 2, 0, time.UTC),
+		time.Date(2026, 6, 4, 15, 0, 3, 0, time.UTC),
+	})})
+	session, err := store.Create(CreateInput{SessionID: "batch"})
+	if err != nil {
+		t.Fatalf("Create returned error: %v", err)
+	}
+
+	appended, err := store.AppendEvents(session.SessionID, []AppendEventInput{
+		{Type: EventMessage, Payload: map[string]any{"role": "user", "content": "one"}},
+		{Type: EventToolCall, Payload: map[string]any{"id": "call_1", "name": "read_file"}},
+		{Type: EventToolResult, Payload: map[string]any{"toolCallId": "call_1", "status": "ok"}},
+	})
+	if err != nil {
+		t.Fatalf("AppendEvents returned error: %v", err)
+	}
+	if len(appended) != 3 {
+		t.Fatalf("expected 3 appended events, got %#v", appended)
+	}
+	for index, event := range appended {
+		wantSequence := index + 1
+		if event.Sequence != wantSequence || event.ID != fmt.Sprintf("batch:%d", wantSequence) {
+			t.Fatalf("event identity mismatch at %d: %#v", index, event)
+		}
+	}
+	if appended[0].CreatedAt != "2026-06-04T15:00:01Z" || appended[2].CreatedAt != "2026-06-04T15:00:03Z" {
+		t.Fatalf("unexpected event timestamps: %#v", appended)
+	}
+
+	loaded, err := store.Get(session.SessionID)
+	if err != nil {
+		t.Fatalf("Get returned error: %v", err)
+	}
+	if loaded == nil || loaded.EventCount != 3 || loaded.LastEventType != EventToolResult || loaded.UpdatedAt != appended[2].CreatedAt {
+		t.Fatalf("metadata not updated from final batch event: %#v", loaded)
+	}
+	events, err := store.ReadEvents(session.SessionID)
+	if err != nil {
+		t.Fatalf("ReadEvents returned error: %v", err)
+	}
+	if !reflect.DeepEqual(eventTypesForTest(events), []EventType{EventMessage, EventToolCall, EventToolResult}) {
+		t.Fatalf("unexpected event types: %#v", events)
+	}
+}
+
+func TestStoreAppendEventsEmptyBatchDoesNotRewriteMetadata(t *testing.T) {
+	store := NewStore(StoreOptions{RootDir: t.TempDir(), Now: fixedClock("2026-06-04T15:10:00Z")})
+	session, err := store.Create(CreateInput{SessionID: "empty_batch", Title: "unchanged"})
+	if err != nil {
+		t.Fatalf("Create returned error: %v", err)
+	}
+
+	appended, err := store.AppendEvents(session.SessionID, nil)
+	if err != nil {
+		t.Fatalf("AppendEvents empty returned error: %v", err)
+	}
+	if len(appended) != 0 {
+		t.Fatalf("empty AppendEvents returned %#v", appended)
+	}
+	loaded, err := store.Get(session.SessionID)
+	if err != nil {
+		t.Fatalf("Get returned error: %v", err)
+	}
+	if loaded == nil || !reflect.DeepEqual(*loaded, session) {
+		t.Fatalf("empty AppendEvents rewrote metadata: before=%#v after=%#v", session, loaded)
+	}
+}
+
+func TestStoreAppendEventsInvalidPayloadDoesNotPartiallyAppend(t *testing.T) {
+	store := NewStore(StoreOptions{RootDir: t.TempDir(), Now: fixedClock("2026-06-04T15:20:00Z")})
+	session, err := store.Create(CreateInput{SessionID: "invalid_batch"})
+	if err != nil {
+		t.Fatalf("Create returned error: %v", err)
+	}
+
+	_, err = store.AppendEvents(session.SessionID, []AppendEventInput{
+		{Type: EventMessage, Payload: map[string]any{"content": "valid first"}},
+		{Type: EventMessage, Payload: json.RawMessage(`{"broken"`)},
+	})
+	if err == nil || !strings.Contains(err.Error(), "invalid raw JSON payload") {
+		t.Fatalf("expected invalid raw payload error, got %v", err)
+	}
+	events, err := store.ReadEvents(session.SessionID)
+	if err != nil {
+		t.Fatalf("ReadEvents returned error: %v", err)
+	}
+	if len(events) != 0 {
+		t.Fatalf("invalid batch should not append partial events: %#v", events)
+	}
+	loaded, err := store.Get(session.SessionID)
+	if err != nil {
+		t.Fatalf("Get returned error: %v", err)
+	}
+	if loaded == nil || loaded.EventCount != 0 || loaded.LastEventType != "" {
+		t.Fatalf("invalid batch should not update metadata: %#v", loaded)
+	}
+}
+
+func TestStoreAppendEventsSequencesAfterMetadataLag(t *testing.T) {
+	store := NewStore(StoreOptions{RootDir: t.TempDir(), Now: fixedClock("2026-06-04T15:30:00Z")})
+	session, err := store.Create(CreateInput{SessionID: "lagging"})
+	if err != nil {
+		t.Fatalf("Create returned error: %v", err)
+	}
+	if _, err := store.AppendEvent(session.SessionID, AppendEventInput{Type: EventMessage, Payload: map[string]any{"content": "already durable"}}); err != nil {
+		t.Fatalf("AppendEvent returned error: %v", err)
+	}
+	lagging, err := store.Get(session.SessionID)
+	if err != nil {
+		t.Fatalf("Get returned error: %v", err)
+	}
+	lagging.EventCount = 0
+	lagging.LastEventType = ""
+	if err := store.writeMetadata(*lagging); err != nil {
+		t.Fatalf("writeMetadata returned error: %v", err)
+	}
+
+	appended, err := store.AppendEvents(session.SessionID, []AppendEventInput{
+		{Type: EventToolCall, Payload: map[string]any{"id": "call"}},
+		{Type: EventToolResult, Payload: map[string]any{"toolCallId": "call"}},
+	})
+	if err != nil {
+		t.Fatalf("AppendEvents returned error: %v", err)
+	}
+	if len(appended) != 2 || appended[0].Sequence != 2 || appended[1].Sequence != 3 {
+		t.Fatalf("batch did not sequence after durable log tail: %#v", appended)
+	}
+	loaded, err := store.Get(session.SessionID)
+	if err != nil {
+		t.Fatalf("Get returned error: %v", err)
+	}
+	if loaded == nil || loaded.EventCount != 3 || loaded.LastEventType != EventToolResult {
+		t.Fatalf("metadata not repaired by batch append: %#v", loaded)
+	}
+}
+
+func TestStoreAppendEventsSerializesConcurrentBatches(t *testing.T) {
+	store := NewStore(StoreOptions{RootDir: t.TempDir(), Now: fixedClock("2026-06-04T15:40:00Z")})
+	session, err := store.Create(CreateInput{SessionID: "concurrent_batches"})
+	if err != nil {
+		t.Fatalf("Create returned error: %v", err)
+	}
+
+	const writers = 8
+	const perBatch = 3
+	var wg sync.WaitGroup
+	errs := make(chan error, writers)
+	for writer := 0; writer < writers; writer++ {
+		wg.Add(1)
+		go func(writer int) {
+			defer wg.Done()
+			appended, err := store.AppendEvents(session.SessionID, []AppendEventInput{
+				{Type: EventMessage, Payload: map[string]int{"writer": writer, "index": 0}},
+				{Type: EventMessage, Payload: map[string]int{"writer": writer, "index": 1}},
+				{Type: EventMessage, Payload: map[string]int{"writer": writer, "index": 2}},
+			})
+			if err != nil {
+				errs <- err
+				return
+			}
+			if len(appended) != perBatch {
+				errs <- fmt.Errorf("writer %d appended %d events", writer, len(appended))
+				return
+			}
+			for index := 1; index < len(appended); index++ {
+				if appended[index].Sequence != appended[index-1].Sequence+1 {
+					errs <- fmt.Errorf("writer %d batch was interleaved: %#v", writer, appended)
+					return
+				}
+			}
+			errs <- nil
+		}(writer)
+	}
+	wg.Wait()
+	close(errs)
+	for err := range errs {
+		if err != nil {
+			t.Fatalf("AppendEvents returned error: %v", err)
+		}
+	}
+
+	events, err := store.ReadEvents(session.SessionID)
+	if err != nil {
+		t.Fatalf("ReadEvents returned error: %v", err)
+	}
+	total := writers * perBatch
+	if len(events) != total {
+		t.Fatalf("expected %d events, got %d", total, len(events))
+	}
+	seen := map[int]bool{}
+	for _, event := range events {
+		if seen[event.Sequence] {
+			t.Fatalf("duplicate sequence %d in %#v", event.Sequence, events)
+		}
+		seen[event.Sequence] = true
+	}
+	for sequence := 1; sequence <= total; sequence++ {
+		if !seen[sequence] {
+			t.Fatalf("missing sequence %d in %#v", sequence, events)
+		}
+	}
+	loaded, err := store.Get(session.SessionID)
+	if err != nil {
+		t.Fatalf("Get returned error: %v", err)
+	}
+	if loaded == nil || loaded.EventCount != total || loaded.LastEventType != EventMessage {
+		t.Fatalf("metadata not updated after concurrent batches: %#v", loaded)
+	}
+}
+
+func eventTypesForTest(events []Event) []EventType {
+	types := make([]EventType, 0, len(events))
+	for _, event := range events {
+		types = append(types, event.Type)
+	}
+	return types
+}

internal/sessions/store.go

@@ -153,6 +153,11 @@ type AppendEventInput struct {
 	Payload any
 }
 
+type preparedAppendEvent struct {
+	Type    EventType
+	Payload json.RawMessage
+}
+
 type RecordSpecInput struct {
 	SpecID              string
 	SpecFilePath        string
@@ -426,7 +431,7 @@ func (store *Store) Fork(parentSessionID string, input ForkInput) (Metadata, err
 	if err != nil {
 		return Metadata{}, err
 	}
-	copied := 0
+	copyInputs := []AppendEventInput{}
 	for _, event := range events {
 		// Do NOT copy usage accounting into the fork. It already counted against the
 		// parent, and a usage report that aggregates the parent and the fork would
@@ -435,11 +440,12 @@ func (store *Store) Fork(parentSessionID string, input ForkInput) (Metadata, err
 		if event.Type == EventUsage {
 			continue
 		}
-		if _, err := store.AppendEvent(fork.SessionID, AppendEventInput{Type: event.Type, Payload: event.Payload}); err != nil {
-			return Metadata{}, err
-		}
-		copied++
+		copyInputs = append(copyInputs, AppendEventInput{Type: event.Type, Payload: event.Payload})
 	}
+	if _, err := store.AppendEvents(fork.SessionID, copyInputs); err != nil {
+		return Metadata{}, err
+	}
+	copied := len(copyInputs)
 	// Copy the parent's content-addressed checkpoint blobs into the fork so the
 	// copied EventSessionCheckpoint events resolve to real blobs and a rewind on
 	// the fork can restore file content (otherwise rewind reads missing blobs
@@ -513,19 +519,34 @@ func (store *Store) RecordSpec(sessionID string, input RecordSpecInput) (Metadat
 }
 
 func (store *Store) AppendEvent(sessionID string, input AppendEventInput) (Event, error) {
-	if !ValidSessionID(sessionID) {
-		return Event{}, fmt.Errorf("invalid zero session id %q", sessionID)
-	}
-	if strings.TrimSpace(string(input.Type)) == "" {
-		return Event{}, fmt.Errorf("zero session event type is required")
-	}
-	unlock, err := store.lockSession(sessionID)
+	events, err := store.AppendEvents(sessionID, []AppendEventInput{input})
 	if err != nil {
 		return Event{}, err
 	}
+	return events[0], nil
+}
+
+// AppendEvents appends a batch of events atomically under one session lock and
+// durability pass. The returned events have contiguous sequences in input order.
+// An empty batch is a no-op and does not rewrite metadata.
+func (store *Store) AppendEvents(sessionID string, inputs []AppendEventInput) ([]Event, error) {
+	if !ValidSessionID(sessionID) {
+		return nil, fmt.Errorf("invalid zero session id %q", sessionID)
+	}
+	prepared, err := prepareAppendEventInputs(inputs)
+	if err != nil {
+		return nil, err
+	}
+	if len(prepared) == 0 {
+		return []Event{}, nil
+	}
+	unlock, err := store.lockSession(sessionID)
+	if err != nil {
+		return nil, err
+	}
 	defer unlock()
 
-	return store.appendEventLocked(sessionID, input)
+	return store.appendPreparedEventsLocked(sessionID, prepared)
 }
 
 // AppendEventUnlessExists appends input only when exists(currentEvents) is false,
@@ -572,14 +593,27 @@ func (store *Store) AppendEventUnlessExists(sessionID string, input AppendEventI
 // the single lock they already hold, instead of re-locking (which would deadlock
 // on the non-reentrant in-process mutex).
 func (store *Store) appendEventLocked(sessionID string, input AppendEventInput) (Event, error) {
-	session, err := store.readMetadata(sessionID)
+	prepared, err := prepareAppendEventInputs([]AppendEventInput{input})
 	if err != nil {
 		return Event{}, err
 	}
-	payload, err := rawPayload(input.Payload)
+	events, err := store.appendPreparedEventsLocked(sessionID, prepared)
 	if err != nil {
 		return Event{}, err
 	}
+	return events[0], nil
+}
+
+// appendPreparedEventsLocked appends pre-validated events WITHOUT acquiring the
+// session lock. The caller MUST already hold store.lockSession(sessionID).
+func (store *Store) appendPreparedEventsLocked(sessionID string, inputs []preparedAppendEvent) ([]Event, error) {
+	if len(inputs) == 0 {
+		return []Event{}, nil
+	}
+	session, err := store.readMetadata(sessionID)
+	if err != nil {
+		return nil, err
+	}
 	sequence := session.EventCount + 1
 	// The events log is the source of truth for sequencing. metadata.EventCount is
 	// persisted separately (after the event line below), so a crash between the two
@@ -590,46 +624,54 @@ func (store *Store) appendEventLocked(sessionID string, input AppendEventInput)
 	if logSeq, err := store.lastEventSequence(sessionID); err == nil && logSeq >= sequence {
 		sequence = logSeq + 1
 	}
-	timestamp := store.timestamp()
-	event := Event{
-		ID:        fmt.Sprintf("%s:%d", sessionID, sequence),
-		SessionID: sessionID,
-		Sequence:  sequence,
-		Type:      input.Type,
-		CreatedAt: timestamp,
-		Payload:   payload,
-	}
-	data, err := json.Marshal(event)
-	if err != nil {
-		return Event{}, fmt.Errorf("encode zero session event: %w", err)
+	events := make([]Event, 0, len(inputs))
+	var batch bytes.Buffer
+	for index, input := range inputs {
+		timestamp := store.timestamp()
+		event := Event{
+			ID:        fmt.Sprintf("%s:%d", sessionID, sequence+index),
+			SessionID: sessionID,
+			Sequence:  sequence + index,
+			Type:      input.Type,
+			CreatedAt: timestamp,
+			Payload:   input.Payload,
+		}
+		data, err := json.Marshal(event)
+		if err != nil {
+			return nil, fmt.Errorf("encode zero session event: %w", err)
+		}
+		batch.Write(data)
+		batch.WriteByte('\n')
+		events = append(events, event)
 	}
 	file, err := os.OpenFile(store.eventsPath(sessionID), os.O_WRONLY|os.O_APPEND|os.O_CREATE, 0o600)
 	if err != nil {
-		return Event{}, fmt.Errorf("append zero session event: %w", err)
+		return nil, fmt.Errorf("append zero session event: %w", err)
 	}
-	if _, err := file.Write(append(data, '\n')); err != nil {
+	if _, err := file.Write(batch.Bytes()); err != nil {
 		_ = file.Close()
-		return Event{}, fmt.Errorf("append zero session event: %w", err)
+		return nil, fmt.Errorf("append zero session event: %w", err)
 	}
-	// fsync the event before reporting success: the derived metadata.json IS
+	// fsync the events before reporting success: the derived metadata.json IS
 	// fsync'd (writeMetadata), so without this a crash after the metadata flush
 	// but before the events.jsonl page reaches disk leaves EventCount ahead of the
-	// durable log — silently losing the just-appended event (incl. the checkpoint
+	// durable log — silently losing just-appended events (incl. checkpoints
 	// that /rewind targets). Make the log at least as durable as its metadata. (AUDIT-M12)
 	if err := file.Sync(); err != nil {
 		_ = file.Close()
-		return Event{}, fmt.Errorf("sync zero session event: %w", err)
+		return nil, fmt.Errorf("sync zero session event: %w", err)
 	}
 	if err := file.Close(); err != nil {
-		return Event{}, fmt.Errorf("close zero session event file: %w", err)
+		return nil, fmt.Errorf("close zero session event file: %w", err)
 	}
-	session.UpdatedAt = timestamp
-	session.EventCount = sequence
-	session.LastEventType = input.Type
+	last := events[len(events)-1]
+	session.UpdatedAt = last.CreatedAt
+	session.EventCount = last.Sequence
+	session.LastEventType = last.Type
 	if err := store.writeMetadata(session); err != nil {
-		return Event{}, err
+		return nil, err
 	}
-	return event, nil
+	return events, nil
 }
 
 // UpdateTitle replaces a session's Title and returns the updated metadata. It is
@@ -964,6 +1006,24 @@ func rawPayload(payload any) (json.RawMessage, error) {
 	return data, nil
 }
 
+func prepareAppendEventInputs(inputs []AppendEventInput) ([]preparedAppendEvent, error) {
+	prepared := make([]preparedAppendEvent, 0, len(inputs))
+	for _, input := range inputs {
+		if strings.TrimSpace(string(input.Type)) == "" {
+			return nil, fmt.Errorf("zero session event type is required")
+		}
+		payload, err := rawPayload(input.Payload)
+		if err != nil {
+			return nil, err
+		}
+		prepared = append(prepared, preparedAppendEvent{
+			Type:    input.Type,
+			Payload: payload,
+		})
+	}
+	return prepared, nil
+}
+
 func envValue(env map[string]string, key string) string {
 	if env != nil {
 		return env[key]

internal/streamjson/streamjson.go

@@ -21,6 +21,7 @@ type EventType string
 const (
 	EventRunStart           EventType = "run_start"
 	EventText               EventType = "text"
+	EventReasoning          EventType = "reasoning"
 	EventToolCall           EventType = "tool_call"
 	EventPermission         EventType = "permission"
 	EventPermissionRequest  EventType = "permission_request"

internal/tools/bash.go

@@ -83,7 +83,12 @@ func (tool bashTool) run(ctx context.Context, args map[string]any, engine *zeroS
 	if err != nil {
 		return errorResult("Error: Invalid arguments for bash: " + err.Error())
 	}
-	if issue := detectShellCommandIssue(commandText, runtime.GOOS); issue != nil {
+	// Resolve the command engine before the MSYS preflight check so an
+	// approved require_escalated call (commandEngine == nil, truly
+	// unsandboxed) can actually bypass the MSYS guard instead of being
+	// hard-blocked by the same check it was meant to escalate past.
+	commandEngine := commandEngineForSandboxPermissions(engine, sandboxPermissions)
+	if issue := detectShellCommandIssue(commandText, runtime.GOOS); issue != nil && !msysGuardBypassed(issue, commandEngine) {
 		return shellIssueBlockResult(*issue)
 	}
 
@@ -106,7 +111,6 @@ func (tool bashTool) run(ctx context.Context, args map[string]any, engine *zeroS
 		"cwd":        relativeCwd,
 		"timeout_ms": strconv.Itoa(timeoutMS),
 	}
-	commandEngine := commandEngineForSandboxPermissions(engine, sandboxPermissions)
 	if commandEngine == nil && sandboxPermissions == SandboxPermissionsRequireEscalated {
 		meta["sandbox_permissions"] = string(SandboxPermissionsRequireEscalated)
 	}
@@ -126,8 +130,8 @@ func (tool bashTool) run(ctx context.Context, args map[string]any, engine *zeroS
 	// can't grow Zero's memory before truncation: only the head+tail each stream
 	// will ever surface to the model are retained, the middle is discarded as it
 	// streams, and the true size is counted for the truncation marker.
-	stdout := newBoundedBuffer(bashOutputBudgetBytes, bashOutputBudgetBytes)
-	stderr := newBoundedBuffer(bashOutputBudgetBytes, bashOutputBudgetBytes)
+	stdout := newBoundedBuffer(bashCaptureBudgetBytes, bashCaptureBudgetBytes)
+	stderr := newBoundedBuffer(bashCaptureBudgetBytes, bashCaptureBudgetBytes)
 	command.Stdout = stdout
 	command.Stderr = stderr
 
@@ -167,7 +171,7 @@ func (tool bashTool) run(ctx context.Context, args map[string]any, engine *zeroS
 		outText, errText, truncated := budgetBashCapture(stdoutText, stdout.total, stderrText, stderrTotal, meta)
 		return Result{
 			Status:    StatusError,
-			Output:    formatBashOutputWithShellHint(commandText, outText, errText, exitCode, meta),
+			Output:    formatBashOutputWithShellHint(outText, errText, exitCode, meta),
 			Truncated: truncated,
 			Meta:      meta,
 		}
@@ -178,7 +182,7 @@ func (tool bashTool) run(ctx context.Context, args map[string]any, engine *zeroS
 	if meta[SandboxLikelyDeniedMeta] == "true" {
 		return Result{
 			Status:    StatusError,
-			Output:    formatBashOutputWithShellHint(commandText, outText, errText, exitCode, meta),
+			Output:    formatBashOutputWithShellHint(outText, errText, exitCode, meta),
 			Truncated: truncated,
 			Meta:      meta,
 		}
@@ -198,6 +202,18 @@ func commandEngineForSandboxPermissions(engine *zeroSandbox.Engine, sandboxPermi
 	return engine
 }
 
+// msysGuardBypassed reports whether a windows_msys_sandbox preflight issue no
+// longer applies because sandbox_permissions: require_escalated actually
+// resolved to unsandboxed, host-level execution (commandEngine is nil only
+// then, per commandEngineForSandboxPermissions). MSYS coreutils fail because
+// of the sandbox's restricted token/handles, so running outside it removes
+// the failure mode this guard exists for. Any other issue kind (e.g. a real
+// cmd.exe syntax problem) still blocks regardless of escalation, since
+// running outside the sandbox does not fix a syntax error.
+func msysGuardBypassed(issue *shellIssue, commandEngine *zeroSandbox.Engine) bool {
+	return issue.Kind == windowsMsysSandboxKind && commandEngine == nil
+}
+
 // appendSandboxBlocks appends a <sandbox_blocks> block listing the denials
 // the sandbox log monitor captured, so the model can see what was blocked. With no
 // blocks the stderr is returned unchanged.
@@ -370,12 +386,21 @@ func formatBashOutput(stdout string, stderr string, exitCode int) string {
 	return strings.Join(parts, "\n")
 }
 
-// bashOutputBudgetBytes caps each of stdout/stderr shown to the model. bash is the
-// one tool that can emit unbounded output (`cat large.log`, `find /`, verbose test
-// runs); every other read/search tool already budgets its output. Head+tail
-// truncation keeps both the start and the end of an oversized stream, since
-// build/test failures usually surface at the tail.
-const bashOutputBudgetBytes = 96 * 1024
+// bashOutputBudgetBytes caps each of stdout/stderr shown to the model (32 KiB
+// ≈ 8k tokens per stream). bash is the one tool that can emit unbounded output
+// (`cat large.log`, `find /`, verbose test runs); a chatty build log at the old
+// 96 KiB budget injected ~24k tokens per call and was then re-billed on every
+// following turn. Head+tail truncation keeps both the start and the end of an
+// oversized stream, since build/test failures usually surface at the tail, and
+// the full captured text is spilled to disk so nothing is lost — the model
+// greps/reads the spill instead of re-running the command.
+const bashOutputBudgetBytes = 32 * 1024
+
+// bashCaptureBudgetBytes bounds the head and tail retained IN MEMORY per
+// stream during capture. Deliberately larger than the emit budget: the extra
+// retained bytes never reach the transcript, but they make the spill file —
+// the model's recovery path — cover 6× more of the output.
+const bashCaptureBudgetBytes = 96 * 1024
 
 // budgetBashOutput truncates stdout and stderr to bashOutputBudgetBytes each,
 // keeping the head and tail of anything larger, and records raw/emitted byte
@@ -395,6 +420,19 @@ func budgetBashCapture(out string, outTotal int, errStr string, errTotal int, me
 	outText, outRaw, outTrunc := truncateHeadTailWithTotal(out, outTotal, bashOutputBudgetBytes)
 	errText, errRaw, errTrunc := truncateHeadTailWithTotal(errStr, errTotal, bashOutputBudgetBytes)
 	truncated := outTrunc || errTrunc
+	if truncated {
+		if spillPath := spillBashStreams(out, outTotal, errStr, errTotal); spillPath != "" {
+			hint := "\n[zero] captured output saved to " + spillPath + " (grep or read_file it instead of re-running)"
+			if errTrunc {
+				errText += hint
+			} else {
+				outText += hint
+			}
+			if meta != nil {
+				meta["spill_path"] = spillPath
+			}
+		}
+	}
 	if meta != nil {
 		emitted := len(outText) + len(errText)
 		meta["raw_bytes"] = strconv.Itoa(outRaw + errRaw)
@@ -407,6 +445,43 @@ func budgetBashCapture(out string, outTotal int, errStr string, errTotal int, me
 	return outText, errText, truncated
 }
 
+// spillBashStreams writes the retained (capture-bounded) stdout and stderr to
+// the spill directory as one sectioned file and returns its path, or "" when
+// spilling fails. The spill holds up to bashCaptureBudgetBytes of head and
+// tail per stream — everything zero kept in memory, which is more than the
+// emit budget shows the model but not necessarily the whole output. When the
+// capture itself dropped the middle of a stream (total exceeds the retained
+// bytes), a gap marker is inserted at the head/tail junction so the spilled
+// log never reads as contiguous when it is not.
+func spillBashStreams(stdout string, stdoutTotal int, stderr string, stderrTotal int) string {
+	stdout = sectionWithCaptureGap(stdout, stdoutTotal)
+	stderr = sectionWithCaptureGap(stderr, stderrTotal)
+	var combined strings.Builder
+	combined.Grow(len(stdout) + len(stderr) + 64)
+	combined.WriteString("### stdout\n")
+	combined.WriteString(stdout)
+	if !strings.HasSuffix(stdout, "\n") {
+		combined.WriteString("\n")
+	}
+	combined.WriteString("### stderr\n")
+	combined.WriteString(stderr)
+	return spillTruncatedOutput("bash", combined.String())
+}
+
+// sectionWithCaptureGap marks the point where boundedBuffer dropped the middle
+// of a stream. When total exceeds the retained bytes, the retained text is the
+// frozen head (bashCaptureBudgetBytes, always full once overflow happened)
+// followed immediately by the rolling tail — the junction sits at the head cap,
+// snapped back to a rune boundary.
+func sectionWithCaptureGap(text string, total int) string {
+	if total <= len(text) || len(text) <= bashCaptureBudgetBytes {
+		return text
+	}
+	head := utf8Prefix(text, bashCaptureBudgetBytes)
+	marker := fmt.Sprintf("\n[zero] capture gap: %d bytes omitted from the middle of this stream\n", total-len(text))
+	return head + marker + text[len(head):]
+}
+
 // boundedBuffer is an io.Writer that retains at most headCap bytes from the start
 // and tailCap bytes from the end of a stream while counting the total written, so
 // a command emitting unbounded output (`cat huge.log`, `yes`) cannot grow Zero's
@@ -461,14 +536,6 @@ func (b *boundedBuffer) retained() string {
 	return string(b.head) + string(b.tail)
 }
 
-// truncateHeadTail keeps the first and last halves of value when it exceeds
-// maxBytes, dropping the middle behind a marker. Returns the possibly-truncated
-// text, the raw byte length, and whether truncation happened.
-func truncateHeadTail(value string, maxBytes int) (string, int, bool) {
-	// value holds the whole stream, so its length is the true raw size.
-	return truncateHeadTailWithTotal(value, len(value), maxBytes)
-}
-
 // truncateHeadTailWithTotal head+tail-truncates value to maxBytes, using total —
 // the full original byte count — for the "N bytes omitted" marker and the raw
 // count. total may exceed len(value) when the middle was already discarded during
@@ -488,9 +555,9 @@ func truncateHeadTailWithTotal(value string, total, maxBytes int) (string, int,
 	return utf8Prefix(value, head) + marker + utf8Suffix(value, tail), total, true
 }
 
-func formatBashOutputWithShellHint(command string, stdout string, stderr string, exitCode int, meta map[string]string) string {
+func formatBashOutputWithShellHint(stdout string, stderr string, exitCode int, meta map[string]string) string {
 	output := formatBashOutput(stdout, stderr, exitCode)
-	if issue := detectShellOutputIssue(command, stdout+"\n"+stderr, runtime.GOOS); issue != nil {
+	if issue := detectShellOutputIssue(stdout+"\n"+stderr, runtime.GOOS); issue != nil {
 		meta["shell_issue"] = issue.Kind
 		output = appendShellIssueHint(output, *issue)
 	}

internal/tools/bash_budget_test.go

@@ -2,11 +2,23 @@ package tools
 
 import (
 	"fmt"
+	"os"
 	"strconv"
 	"strings"
 	"testing"
 )
 
+// setTestTempDir points every platform's temp-dir lookup at a per-test
+// directory so spill files never land in the real temp dir: os.TempDir reads
+// TMPDIR on Unix but TMP/TEMP on Windows.
+func setTestTempDir(t *testing.T) {
+	t.Helper()
+	dir := t.TempDir()
+	t.Setenv("TMPDIR", dir)
+	t.Setenv("TMP", dir)
+	t.Setenv("TEMP", dir)
+}
+
 // Small output passes through untouched and records raw==emitted, no truncated flag.
 func TestBudgetBashOutputSmallPassesThrough(t *testing.T) {
 	meta := map[string]string{}
@@ -26,8 +38,10 @@ func TestBudgetBashOutputSmallPassesThrough(t *testing.T) {
 }
 
 // Oversized stdout is truncated head+tail: both the first and last lines survive,
-// the middle is dropped behind a marker, and meta is flagged.
+// the middle is dropped behind a marker, meta is flagged, and the captured text
+// is spilled to a re-readable file.
 func TestBudgetBashOutputTruncatesHeadAndTail(t *testing.T) {
+	setTestTempDir(t)
 	head := "FIRST_LINE_MARKER\n"
 	tail := "\nLAST_LINE_MARKER"
 	big := head + strings.Repeat("x", bashOutputBudgetBytes) + tail
@@ -47,9 +61,23 @@ func TestBudgetBashOutputTruncatesHeadAndTail(t *testing.T) {
 	if !strings.Contains(out, "output truncated") {
 		t.Fatalf("expected a truncation marker, got:\n%s", out[:200])
 	}
-	if len(out) > bashOutputBudgetBytes {
+	// The spill hint (path length varies) is the only allowed overage past the budget.
+	if len(out) > bashOutputBudgetBytes+512 {
 		t.Fatalf("emitted %d bytes exceeds budget %d", len(out), bashOutputBudgetBytes)
 	}
+	if !strings.Contains(out, "captured output saved to ") {
+		t.Fatalf("truncated bash output must carry a spill hint:\n%s", out[len(out)-300:])
+	}
+	if meta["spill_path"] == "" {
+		t.Fatal("spill path missing from meta")
+	}
+	content, err := os.ReadFile(meta["spill_path"])
+	if err != nil {
+		t.Fatalf("spill file unreadable: %v", err)
+	}
+	if !strings.Contains(string(content), "### stdout") || !strings.Contains(string(content), strings.Repeat("x", 1024)) {
+		t.Fatalf("spill must hold the sectioned captured streams (got %d bytes)", len(content))
+	}
 	if meta["truncated"] != "true" {
 		t.Fatalf("expected truncated=true, got %v", meta)
 	}
@@ -100,10 +128,12 @@ func TestBoundedBufferKeepsHeadAndTailBounded(t *testing.T) {
 // budgetBashCapture reports the TRUE total (not the retained size) in the marker
 // and raw_bytes, even though only a bounded head+tail was ever held in memory.
 func TestBudgetBashCaptureReportsTrueTotal(t *testing.T) {
-	// Retained head+tail as boundedBuffer would hand over; the real command produced
-	// far more than was kept.
-	retained := "HEAD_START" + strings.Repeat("y", bashOutputBudgetBytes) + "TAIL_END"
-	total := 10 * bashOutputBudgetBytes
+	setTestTempDir(t)
+	// Retained head+tail as boundedBuffer would hand over after overflow: the
+	// frozen head is full at bashCaptureBudgetBytes, then the rolling tail; the
+	// real command produced far more than was kept.
+	retained := "HEAD_START" + strings.Repeat("y", bashCaptureBudgetBytes) + "TAIL_END"
+	total := 10 * bashCaptureBudgetBytes
 
 	meta := map[string]string{}
 	out, _, truncated := budgetBashCapture(retained, total, "", 0, meta)
@@ -114,7 +144,8 @@ func TestBudgetBashCaptureReportsTrueTotal(t *testing.T) {
 	if !strings.Contains(out, "HEAD_START") || !strings.Contains(out, "TAIL_END") {
 		t.Fatalf("head/tail lost after budgeting:\n%s", out[:min(120, len(out))])
 	}
-	if len(out) > bashOutputBudgetBytes {
+	// Spill hint is the only allowed overage past the budget.
+	if len(out) > bashOutputBudgetBytes+512 {
 		t.Fatalf("emitted %d bytes exceeds budget %d", len(out), bashOutputBudgetBytes)
 	}
 	if meta["raw_bytes"] != strconv.Itoa(total) {
@@ -124,4 +155,29 @@ func TestBudgetBashCaptureReportsTrueTotal(t *testing.T) {
 	if !strings.Contains(out, strconv.Itoa(total-bashOutputBudgetBytes)) {
 		t.Fatalf("marker should cite the true omitted byte count:\n%s", out[:min(200, len(out))])
 	}
+
+	// The spill must be recorded, sectioned, and carry a capture-gap marker at
+	// the head/tail junction — the capture dropped the middle, so the spilled
+	// log must never read as contiguous.
+	if meta["spill_path"] == "" {
+		t.Fatal("spill path missing from meta")
+	}
+	content, err := os.ReadFile(meta["spill_path"])
+	if err != nil {
+		t.Fatalf("spill file unreadable: %v", err)
+	}
+	if !strings.Contains(string(content), "### stdout") {
+		t.Fatal("spill must be sectioned by stream")
+	}
+	gapMarker := fmt.Sprintf("capture gap: %d bytes omitted", total-len(retained))
+	gapIndex := strings.Index(string(content), gapMarker)
+	if gapIndex < 0 {
+		t.Fatalf("spill missing capture-gap marker %q", gapMarker)
+	}
+	// The junction sits right after the frozen head (plus the section header),
+	// not appended at the end of the stream.
+	headerOffset := len("### stdout\n")
+	if gapIndex < headerOffset+bashCaptureBudgetBytes-64 || gapIndex > headerOffset+bashCaptureBudgetBytes+64 {
+		t.Fatalf("capture-gap marker at offset %d, want ~%d (the head cap)", gapIndex, headerOffset+bashCaptureBudgetBytes)
+	}
 }

internal/tools/bash_tool_test.go

@@ -181,9 +181,13 @@ func TestDetectShellCommandIssueFlagsPipedPosixUtilities(t *testing.T) {
 		`cat file.txt | wc -l`,
 		`some_command | tail -n 20`,
 	} {
-		if issue := detectShellCommandIssue(command, "windows"); issue == nil {
+		issue := detectShellCommandIssue(command, "windows")
+		if issue == nil {
 			t.Fatalf("expected POSIX utility pipeline to be flagged for %q", command)
 		}
+		if issue.Kind != "windows_msys_sandbox" {
+			t.Fatalf("expected windows_msys_sandbox for %q, got %q", command, issue.Kind)
+		}
 	}
 }
 
@@ -195,7 +199,7 @@ func TestDetectShellCommandIssueAllowsUnrelatedCommands(t *testing.T) {
 		`echo header text`,
 		`tail-log.sh`,
 		`grep-cli --version`,
-		`sed.exe --help`,
+		`mytool.exe --help`,
 	} {
 		if issue := detectShellCommandIssue(command, "windows"); issue != nil {
 			t.Fatalf("expected unrelated command to pass for %q, got %#v", command, issue)
@@ -204,7 +208,7 @@ func TestDetectShellCommandIssueAllowsUnrelatedCommands(t *testing.T) {
 }
 
 func TestDetectShellOutputIssueAddsWindowsSyntaxHint(t *testing.T) {
-	issue := detectShellOutputIssue(`cd /d/tmp/zero-pr-158 && ls -la`, "The syntax of the command is incorrect.", "windows")
+	issue := detectShellOutputIssue("The syntax of the command is incorrect.", "windows")
 	if issue == nil {
 		t.Fatal("expected Windows syntax error to get shell guidance")
 	}
@@ -434,6 +438,105 @@ func TestRegistryRunsWithDegradedUnavailableNativeSandbox(t *testing.T) {
 	}
 }
 
+// TestBashToolRequireEscalatedMsysGuard covers the boundary of the fix for an
+// escalation-vs-preflight ordering bug: the MSYS sandbox guard exists only
+// because MSYS/Cygwin coreutils fail under the write-restricted sandbox, so
+// once sandbox_permissions: require_escalated is actually approved (running
+// the command unsandboxed), the guard must get out of the way, but an
+// unrelated windows_shell_syntax issue (a real cmd.exe syntax problem, not a
+// sandbox interaction) must still block regardless of escalation.
+func TestBashToolRequireEscalatedMsysGuard(t *testing.T) {
+	if runtime.GOOS != "windows" {
+		t.Skip("windows-only MSYS sandbox guard")
+	}
+	root := t.TempDir()
+	newEngine := func() *sandbox.Engine {
+		return sandbox.NewEngine(sandbox.EngineOptions{
+			WorkspaceRoot: root,
+			Policy:        sandbox.DefaultPolicy(),
+			Backend:       sandbox.Backend{Name: sandbox.BackendUnavailable, Message: "native sandbox unavailable"},
+		})
+	}
+	registry := NewRegistry()
+	registry.Register(NewBashTool(root))
+
+	t.Run("default sandboxing still blocks an MSYS-prone command", func(t *testing.T) {
+		result := registry.RunWithOptions(context.Background(), "bash", map[string]any{
+			"command": "cat somefile.txt",
+		}, RunOptions{
+			PermissionGranted: true,
+			Sandbox:           newEngine(),
+			PermissionMode:    string(sandbox.PermissionModeAsk),
+		})
+		if result.Meta["shell_issue"] != "windows_msys_sandbox" {
+			t.Fatalf("expected windows_msys_sandbox block without escalation, got %#v", result)
+		}
+	})
+
+	t.Run("approved require_escalated bypasses the MSYS guard", func(t *testing.T) {
+		result := registry.RunWithOptions(context.Background(), "bash", map[string]any{
+			"command":             "cat somefile.txt",
+			"sandbox_permissions": string(SandboxPermissionsRequireEscalated),
+		}, RunOptions{
+			PermissionGranted: true,
+			Sandbox:           newEngine(),
+			PermissionMode:    string(sandbox.PermissionModeAsk),
+		})
+		// Assert on the preflight block sentinel (exit_code "-1", set only by
+		// shellIssueBlockResult) rather than shell_issue: once the guard is
+		// bypassed, "cat somefile.txt" actually runs, and its real,
+		// PATH-dependent output could otherwise trip the unrelated
+		// post-execution detectShellOutputIssue heuristic and make this
+		// assertion flaky for reasons unrelated to the guard under test.
+		if result.Meta["exit_code"] == "-1" {
+			t.Fatalf("expected approved require_escalated to bypass the MSYS guard, got blocked: %#v", result)
+		}
+	})
+
+	t.Run("approved require_escalated still blocks an unrelated syntax issue", func(t *testing.T) {
+		result := registry.RunWithOptions(context.Background(), "bash", map[string]any{
+			"command":             `cd /d/tmp/zero-pr-158 && dir`,
+			"sandbox_permissions": string(SandboxPermissionsRequireEscalated),
+		}, RunOptions{
+			PermissionGranted: true,
+			Sandbox:           newEngine(),
+			PermissionMode:    string(sandbox.PermissionModeAsk),
+		})
+		if result.Meta["shell_issue"] != "windows_shell_syntax" {
+			t.Fatalf("expected windows_shell_syntax block to still apply under require_escalated, got %#v", result)
+		}
+	})
+}
+
+// TestBashToolIgnoresMsysMarkersInCommandArgumentsAfterFailure guards the
+// post-execution MSYS heuristic against treating the command line itself as
+// evidence. The helper command below fails for a reason unrelated to MSYS
+// (a plain non-zero exit), but its own argument text quotes an MSYS crash
+// marker the way a PR-comment or commit-message argument might. Before the
+// fix, detectShellOutputIssue scanned command+output together and would have
+// misdiagnosed this as an MSYS sandbox failure even though the real
+// stdout/stderr never mentions MSYS at all.
+func TestBashToolIgnoresMsysMarkersInCommandArgumentsAfterFailure(t *testing.T) {
+	if runtime.GOOS != "windows" {
+		t.Skip("windows-only MSYS output guard")
+	}
+	root := t.TempDir()
+	// The helper only reads argv[2] ("fail"); this trailing quoted argument is
+	// otherwise unused by the helper but still part of the command line text.
+	command := helperCommand("fail") + ` "fatal error - CreateFileMapping S-1-5-21, Win32 error 5. Terminating. cygheap_user::init"`
+
+	result := NewBashTool(root).Run(context.Background(), map[string]any{
+		"command": command,
+	})
+
+	if result.Status != StatusError || result.Meta["exit_code"] != "7" {
+		t.Fatalf("expected the helper's real exit 7 failure, got %s: %#v", result.Status, result)
+	}
+	if result.Meta["shell_issue"] == "windows_msys_sandbox" {
+		t.Fatalf("expected the MSYS marker in the command's own argument text to be ignored, got %#v", result)
+	}
+}
+
 func TestBashToolRunsWithDegradedUnavailableNativeSandbox(t *testing.T) {
 	root := t.TempDir()
 	policy := sandbox.DefaultPolicy()

internal/tools/edit_file.go

@@ -2,6 +2,7 @@ package tools
 
 import (
 	"context"
+	"errors"
 	"fmt"
 	"os"
 	"strings"
@@ -44,7 +45,7 @@ func (tool editFileTool) Run(ctx context.Context, args map[string]any) Result {
 	return tool.RunWithOptions(ctx, args, RunOptions{})
 }
 
-func (tool editFileTool) RunWithOptions(_ context.Context, args map[string]any, options RunOptions) Result {
+func (tool editFileTool) RunWithOptions(ctx context.Context, args map[string]any, options RunOptions) Result {
 	requestedPath, err := aliasedStringArg(args, []string{"path", "file", "file_path", "filename"}, "", true, false)
 	if err != nil {
 		return errorResult("Error: Invalid arguments for edit_file: " + err.Error())
@@ -98,6 +99,35 @@ func (tool editFileTool) RunWithOptions(_ context.Context, args map[string]any,
 		}
 	}
 
+	// Fuzzy fallback: when the exact string (and its CRLF translation) is absent,
+	// run a cascade of tolerant matchers (trimmed lines, block anchors, collapsed
+	// whitespace, indentation drift, escape normalization) to locate the span the
+	// model intended. Only a span that occurs literally in the file is accepted,
+	// so the replacement applied below is still exact.
+	if occurrences == 0 {
+		findOld, findNew := oldString, newString
+		if strings.Contains(content, "\r\n") && !strings.Contains(findOld, "\r\n") {
+			findOld = strings.ReplaceAll(findOld, "\n", "\r\n")
+			findNew = strings.ReplaceAll(findNew, "\n", "\r\n")
+		}
+		search, ferr := fuzzyEditMatch(content, findOld, replaceAll)
+		switch {
+		case ferr == nil:
+			oldString = search
+			// The model's new_string was written at old_string's (mismatched)
+			// indentation; re-shape it to the span actually being replaced so a
+			// tolerant match never strips indentation or a trailing CR.
+			newString = adaptReplacementToSpan(search, findOld, findNew)
+			occurrences = strings.Count(content, search)
+		case errors.Is(ferr, errEditFuzzyAmbiguous):
+			return errorResult("Error: old_string matches multiple locations in " + relativePath + " even after fuzzy matching. Provide more surrounding context to make the match unique, or pass replace_all: true.")
+		case errors.Is(ferr, errEditFuzzyNotFound):
+			// Fall through to the exact-match error below.
+		default:
+			return errorResult("Error editing " + relativePath + ": " + ferr.Error())
+		}
+	}
+
 	if occurrences == 0 {
 		return errorResult("Error: Could not find the exact string to replace in " + relativePath + ". The old_string must match the file byte-for-byte.")
 	}
@@ -120,6 +150,10 @@ func (tool editFileTool) RunWithOptions(_ context.Context, args map[string]any,
 	if err := os.WriteFile(absolutePath, []byte(updated), 0o644); err != nil {
 		return errorResult("Error writing " + relativePath + ": " + err.Error())
 	}
+	// Optional format-on-write (ZERO_FORMAT_ON_WRITE). Must run BEFORE the
+	// FileTracker re-baseline: recording pre-format content would make the very
+	// next edit look like an external modification and trip the conflict guard.
+	updated = maybeFormatWrittenFile(ctx, absolutePath, updated)
 	// Re-baseline to the content we just wrote so subsequent edits in this session
 	// compare against the current on-disk state, not the pre-edit version.
 	newInfo, _ := os.Stat(absolutePath)
@@ -130,6 +164,7 @@ func (tool editFileTool) RunWithOptions(_ context.Context, args map[string]any,
 		suffix = "s"
 	}
 	summary := fmt.Sprintf("Successfully edited %s (replaced %d occurrence%s).", relativePath, replacedCount, suffix)
+	summary += inlineDiagnostics(ctx, options, absolutePath, relativePath)
 	result := okResult(summary)
 	result.ChangedFiles = []string{relativePath}
 	// Card-only preview (Display.Preview): the model's Output stays the one-line

internal/tools/edit_replacers.go

@@ -0,0 +1,554 @@
+package tools
+
+// Fuzzy fallback matching for edit_file. When the model's old_string fails to
+// match byte-for-byte (drifted indentation, collapsed whitespace, escaped
+// characters, a slightly-misremembered middle line), these replacers propose
+// candidate spans that plausibly correspond to what the model intended. Only a
+// candidate that occurs literally in the file is accepted, so the replacement
+// itself is always exact even when the match was tolerant.
+//
+// Strategy cascade ported from opencode's edit tool, whose replacers were in
+// turn distilled from Cline's diff-apply evals and gemini-cli's edit corrector.
+
+import (
+	"errors"
+	"regexp"
+	"sort"
+	"strings"
+)
+
+// editReplacer proposes candidate spans for find inside content. Candidates
+// are re-validated by fuzzyEditMatch before use.
+type editReplacer func(content, find string) []string
+
+var (
+	errEditFuzzyNotFound  = errors.New("no fuzzy match for old_string")
+	errEditFuzzyAmbiguous = errors.New("fuzzy match for old_string is ambiguous")
+)
+
+// Minimum average middle-line similarity for block-anchor matches.
+const editAnchorSimilarityThreshold = 0.65
+
+// fuzzyEditMatch runs the replacer cascade and returns the exact span of
+// content to replace. When replaceAll is false the span must be unique in
+// content; an ambiguous candidate is skipped in favor of later candidates and
+// only reported if nothing unique is found. A span wildly larger than
+// old_string is refused outright rather than risking a destructive edit.
+func fuzzyEditMatch(content, find string, replaceAll bool) (string, error) {
+	replacers := []editReplacer{
+		lineTrimmedReplacer,
+		blockAnchorReplacer,
+		whitespaceNormalizedReplacer,
+		indentationFlexibleReplacer,
+		escapeNormalizedReplacer,
+		trimmedBoundaryReplacer,
+		contextAwareReplacer,
+	}
+	found := false
+	for _, replacer := range replacers {
+		// Collect the replacer's DISTINCT candidate spans that literally occur in
+		// content. Two or more distinct spans from one strategy (e.g. duplicate
+		// blocks at different indentation, each occurring once) mean the model's
+		// intent is genuinely ambiguous — silently editing the first would be a
+		// wrong-span write, so it is rejected instead.
+		var candidates []string
+		seen := map[string]bool{}
+		for _, search := range replacer(content, find) {
+			if search == "" || seen[search] {
+				continue
+			}
+			seen[search] = true
+			if !strings.Contains(content, search) {
+				continue
+			}
+			candidates = append(candidates, search)
+		}
+		if len(candidates) == 0 {
+			continue
+		}
+		found = true
+		if !replaceAll && len(candidates) > 1 {
+			return "", errEditFuzzyAmbiguous
+		}
+		search := candidates[0]
+		if isDisproportionateEditMatch(search, find) {
+			return "", errors.New("refusing replacement because the matched span is much larger than old_string; re-read the file and provide the full exact old_string for the intended replacement")
+		}
+		if replaceAll {
+			return search, nil
+		}
+		if strings.Index(content, search) == strings.LastIndex(content, search) {
+			return search, nil
+		}
+		// The single candidate occurs at multiple positions; a later, stricter
+		// strategy may still resolve a unique span, so keep cascading.
+	}
+	if !found {
+		return "", errEditFuzzyNotFound
+	}
+	return "", errEditFuzzyAmbiguous
+}
+
+// isDisproportionateEditMatch guards against anchor-style replacers matching a
+// span far larger than the text the model asked to replace (e.g. first/last
+// line anchors bridging hundreds of unrelated lines).
+func isDisproportionateEditMatch(search, find string) bool {
+	findLines := strings.Count(find, "\n") + 1
+	searchLines := strings.Count(search, "\n") + 1
+	limit := findLines + 3
+	if findLines*2 > limit {
+		limit = findLines * 2
+	}
+	if searchLines >= limit {
+		return true
+	}
+	if findLines == 1 {
+		return false
+	}
+	searchTrimmed := len(strings.TrimSpace(search))
+	findTrimmed := len(strings.TrimSpace(find))
+	byteLimit := findTrimmed + 500
+	if findTrimmed*4 > byteLimit {
+		byteLimit = findTrimmed * 4
+	}
+	return searchTrimmed > byteLimit
+}
+
+// splitFindLines splits find into lines, dropping the trailing empty line a
+// trailing newline produces so windows align with real content lines.
+func splitFindLines(find string) []string {
+	lines := strings.Split(find, "\n")
+	if len(lines) > 0 && lines[len(lines)-1] == "" {
+		lines = lines[:len(lines)-1]
+	}
+	return lines
+}
+
+// lineTrimmedReplacer matches when every line equals the corresponding
+// old_string line after trimming surrounding whitespace.
+func lineTrimmedReplacer(content, find string) []string {
+	contentLines := strings.Split(content, "\n")
+	findLines := splitFindLines(find)
+	if len(findLines) == 0 {
+		return nil
+	}
+	var candidates []string
+	for i := 0; i+len(findLines) <= len(contentLines); i++ {
+		matches := true
+		for j := range findLines {
+			if strings.TrimSpace(contentLines[i+j]) != strings.TrimSpace(findLines[j]) {
+				matches = false
+				break
+			}
+		}
+		if matches {
+			candidates = append(candidates, strings.Join(contentLines[i:i+len(findLines)], "\n"))
+		}
+	}
+	return candidates
+}
+
+// blockAnchorReplacer anchors on the first and last lines (trimmed) and
+// accepts the block when the middle lines average >= the similarity threshold
+// (Levenshtein-based), tolerating a slightly misremembered interior.
+func blockAnchorReplacer(content, find string) []string {
+	findLines := splitFindLines(find)
+	if len(findLines) < 3 {
+		return nil
+	}
+	contentLines := strings.Split(content, "\n")
+	firstAnchor := strings.TrimSpace(findLines[0])
+	lastAnchor := strings.TrimSpace(findLines[len(findLines)-1])
+	searchBlockSize := len(findLines)
+	maxLineDelta := searchBlockSize / 4
+	if maxLineDelta < 1 {
+		maxLineDelta = 1
+	}
+
+	type span struct{ start, end int }
+	var candidates []span
+	for i := 0; i < len(contentLines); i++ {
+		if strings.TrimSpace(contentLines[i]) != firstAnchor {
+			continue
+		}
+		// Only the first occurrence of the last anchor after this start counts,
+		// mirroring the reference implementation: a farther-away closing line is
+		// assumed to close a different block.
+		for j := i + 2; j < len(contentLines); j++ {
+			if strings.TrimSpace(contentLines[j]) != lastAnchor {
+				continue
+			}
+			actualBlockSize := j - i + 1
+			delta := actualBlockSize - searchBlockSize
+			if delta < 0 {
+				delta = -delta
+			}
+			if delta <= maxLineDelta {
+				candidates = append(candidates, span{start: i, end: j})
+			}
+			break
+		}
+	}
+	if len(candidates) == 0 {
+		return nil
+	}
+
+	middleSimilarity := func(candidate span) float64 {
+		actualBlockSize := candidate.end - candidate.start + 1
+		linesToCheck := searchBlockSize - 2
+		if actualBlockSize-2 < linesToCheck {
+			linesToCheck = actualBlockSize - 2
+		}
+		if linesToCheck <= 0 {
+			return 1.0
+		}
+		similarity := 0.0
+		for j := 1; j < searchBlockSize-1 && j < actualBlockSize-1; j++ {
+			contentLine := strings.TrimSpace(contentLines[candidate.start+j])
+			findLine := strings.TrimSpace(findLines[j])
+			maxLen := len(contentLine)
+			if len(findLine) > maxLen {
+				maxLen = len(findLine)
+			}
+			if maxLen == 0 {
+				continue
+			}
+			distance := levenshtein(contentLine, findLine)
+			similarity += 1 - float64(distance)/float64(maxLen)
+		}
+		return similarity / float64(linesToCheck)
+	}
+
+	// Return EVERY candidate that clears the similarity threshold, best first.
+	// Picking only the best would hide competing blocks from fuzzyEditMatch's
+	// distinct-candidate ambiguity check and could silently edit the wrong
+	// block; with all qualifying spans surfaced, one clear winner still
+	// resolves (single candidate) while two plausible blocks are rejected as
+	// ambiguous. replaceAll consumers take the first (most similar) span.
+	type scoredSpan struct {
+		span       span
+		similarity float64
+	}
+	var qualifying []scoredSpan
+	for _, candidate := range candidates {
+		if s := middleSimilarity(candidate); s >= editAnchorSimilarityThreshold {
+			qualifying = append(qualifying, scoredSpan{span: candidate, similarity: s})
+		}
+	}
+	if len(qualifying) == 0 {
+		return nil
+	}
+	sort.SliceStable(qualifying, func(i, j int) bool {
+		return qualifying[i].similarity > qualifying[j].similarity
+	})
+	spans := make([]string, 0, len(qualifying))
+	for _, scored := range qualifying {
+		spans = append(spans, strings.Join(contentLines[scored.span.start:scored.span.end+1], "\n"))
+	}
+	return spans
+}
+
+var editWhitespaceRun = regexp.MustCompile(`\s+`)
+
+func normalizeEditWhitespace(text string) string {
+	return strings.TrimSpace(editWhitespaceRun.ReplaceAllString(text, " "))
+}
+
+// whitespaceNormalizedReplacer matches after collapsing all whitespace runs to
+// a single space: full lines, sub-line spans (via a word-boundary regex), and
+// multi-line windows.
+func whitespaceNormalizedReplacer(content, find string) []string {
+	normalizedFind := normalizeEditWhitespace(find)
+	if normalizedFind == "" {
+		return nil
+	}
+	contentLines := strings.Split(content, "\n")
+	var candidates []string
+	var subLinePattern *regexp.Regexp
+	for _, line := range contentLines {
+		normalizedLine := normalizeEditWhitespace(line)
+		if normalizedLine == normalizedFind {
+			candidates = append(candidates, line)
+			continue
+		}
+		if !strings.Contains(normalizedLine, normalizedFind) {
+			continue
+		}
+		if subLinePattern == nil {
+			words := strings.Fields(find)
+			quoted := make([]string, len(words))
+			for i, word := range words {
+				quoted[i] = regexp.QuoteMeta(word)
+			}
+			pattern, err := regexp.Compile(strings.Join(quoted, `\s+`))
+			if err != nil {
+				continue
+			}
+			subLinePattern = pattern
+		}
+		if match := subLinePattern.FindString(line); match != "" {
+			candidates = append(candidates, match)
+		}
+	}
+
+	findLines := strings.Split(find, "\n")
+	if len(findLines) > 1 {
+		for i := 0; i+len(findLines) <= len(contentLines); i++ {
+			block := strings.Join(contentLines[i:i+len(findLines)], "\n")
+			if normalizeEditWhitespace(block) == normalizedFind {
+				candidates = append(candidates, block)
+			}
+		}
+	}
+	return candidates
+}
+
+// stripCommonIndentation removes the minimum leading-whitespace width shared
+// by all non-empty lines, so blocks match regardless of their nesting depth.
+func stripCommonIndentation(text string) string {
+	lines := strings.Split(text, "\n")
+	minIndent := -1
+	for _, line := range lines {
+		if strings.TrimSpace(line) == "" {
+			continue
+		}
+		indent := len(line) - len(strings.TrimLeft(line, " \t"))
+		if minIndent < 0 || indent < minIndent {
+			minIndent = indent
+		}
+	}
+	if minIndent <= 0 {
+		return text
+	}
+	stripped := make([]string, len(lines))
+	for i, line := range lines {
+		if strings.TrimSpace(line) == "" {
+			stripped[i] = line
+			continue
+		}
+		stripped[i] = line[minIndent:]
+	}
+	return strings.Join(stripped, "\n")
+}
+
+func indentationFlexibleReplacer(content, find string) []string {
+	normalizedFind := stripCommonIndentation(find)
+	contentLines := strings.Split(content, "\n")
+	findLines := strings.Split(find, "\n")
+	var candidates []string
+	for i := 0; i+len(findLines) <= len(contentLines); i++ {
+		block := strings.Join(contentLines[i:i+len(findLines)], "\n")
+		if stripCommonIndentation(block) == normalizedFind {
+			candidates = append(candidates, block)
+		}
+	}
+	return candidates
+}
+
+var editEscapeSequence = regexp.MustCompile("\\\\(n|t|r|'|\"|`|\\\\|\\n|\\$)")
+
+// unescapeEditString undoes one level of string escaping (\n, \t, \", \\, a
+// backslash-newline continuation, \$) — the model sometimes reproduces file
+// content as it appeared inside a quoted string literal.
+func unescapeEditString(text string) string {
+	return editEscapeSequence.ReplaceAllStringFunc(text, func(match string) string {
+		switch match[1:] {
+		case "n":
+			return "\n"
+		case "t":
+			return "\t"
+		case "r":
+			return "\r"
+		case "\n":
+			return "\n"
+		default:
+			// ', ", `, \, $ all unescape to themselves.
+			return match[1:]
+		}
+	})
+}
+
+func escapeNormalizedReplacer(content, find string) []string {
+	unescapedFind := unescapeEditString(find)
+	var candidates []string
+	if strings.Contains(content, unescapedFind) {
+		candidates = append(candidates, unescapedFind)
+	}
+	contentLines := strings.Split(content, "\n")
+	findLines := strings.Split(unescapedFind, "\n")
+	for i := 0; i+len(findLines) <= len(contentLines); i++ {
+		block := strings.Join(contentLines[i:i+len(findLines)], "\n")
+		if unescapeEditString(block) == unescapedFind {
+			candidates = append(candidates, block)
+		}
+	}
+	return candidates
+}
+
+// trimmedBoundaryReplacer tolerates stray leading/trailing whitespace (often
+// blank lines) around an otherwise-exact old_string.
+func trimmedBoundaryReplacer(content, find string) []string {
+	trimmedFind := strings.TrimSpace(find)
+	if trimmedFind == find || trimmedFind == "" {
+		return nil
+	}
+	var candidates []string
+	if strings.Contains(content, trimmedFind) {
+		candidates = append(candidates, trimmedFind)
+	}
+	contentLines := strings.Split(content, "\n")
+	findLines := strings.Split(find, "\n")
+	for i := 0; i+len(findLines) <= len(contentLines); i++ {
+		block := strings.Join(contentLines[i:i+len(findLines)], "\n")
+		if strings.TrimSpace(block) == trimmedFind {
+			candidates = append(candidates, block)
+		}
+	}
+	return candidates
+}
+
+// contextAwareReplacer anchors on the first and last lines and accepts an
+// equal-length block when at least half of its non-empty middle lines match
+// after trimming — a cheaper, stricter cousin of blockAnchorReplacer.
+func contextAwareReplacer(content, find string) []string {
+	findLines := splitFindLines(find)
+	if len(findLines) < 3 {
+		return nil
+	}
+	contentLines := strings.Split(content, "\n")
+	firstAnchor := strings.TrimSpace(findLines[0])
+	lastAnchor := strings.TrimSpace(findLines[len(findLines)-1])
+	var candidates []string
+	for i := 0; i < len(contentLines); i++ {
+		if strings.TrimSpace(contentLines[i]) != firstAnchor {
+			continue
+		}
+		for j := i + 2; j < len(contentLines); j++ {
+			if strings.TrimSpace(contentLines[j]) != lastAnchor {
+				continue
+			}
+			if j-i+1 == len(findLines) {
+				matching, nonEmpty := 0, 0
+				for k := 1; k < len(findLines)-1; k++ {
+					blockLine := strings.TrimSpace(contentLines[i+k])
+					findLine := strings.TrimSpace(findLines[k])
+					if blockLine == "" && findLine == "" {
+						continue
+					}
+					nonEmpty++
+					if blockLine == findLine {
+						matching++
+					}
+				}
+				if nonEmpty == 0 || float64(matching)/float64(nonEmpty) >= 0.5 {
+					candidates = append(candidates, strings.Join(contentLines[i:j+1], "\n"))
+				}
+			}
+			break
+		}
+	}
+	return candidates
+}
+
+// adaptReplacementToSpan re-shapes the model's replacement to the span a
+// tolerant matcher resolved. When old_string only matched after normalization,
+// new_string was written at old_string's (wrong) shape, so applying it raw
+// would strip the file's indentation or drop a trailing CR:
+//
+//  1. Uniform re-indent: when every span line equals delta + the corresponding
+//     find line's indentation (the line-trimmed / indentation-flexible shapes),
+//     the same delta is prepended to every non-blank replacement line. Any
+//     line that breaks the uniform-delta relationship disables the shift —
+//     block-anchor matches with a drifted interior are left untouched.
+//  2. Trailing CR: a span from a CRLF file ends mid-line at "\r" (candidates
+//     are built by joining lines split on "\n"); the replacement gets the same
+//     trailing "\r" so the file's CRLF pairs stay intact.
+func adaptReplacementToSpan(span, find, replacement string) string {
+	if delta, ok := uniformIndentDelta(span, find); ok && delta != "" {
+		lines := strings.Split(replacement, "\n")
+		for i, line := range lines {
+			if strings.TrimSpace(line) == "" {
+				continue
+			}
+			lines[i] = delta + line
+		}
+		replacement = strings.Join(lines, "\n")
+	}
+	if strings.HasSuffix(span, "\r") && !strings.HasSuffix(replacement, "\r") {
+		replacement += "\r"
+	}
+	return replacement
+}
+
+// uniformIndentDelta returns the indentation prefix that, prepended to every
+// non-blank find line, yields the corresponding span line's indentation. ok is
+// false when line counts differ, any line pair disagrees on the delta, or the
+// span is not simply a uniformly deeper-indented copy of find.
+func uniformIndentDelta(span, find string) (string, bool) {
+	spanLines := strings.Split(span, "\n")
+	findLines := splitFindLines(find)
+	if len(spanLines) != len(findLines) {
+		return "", false
+	}
+	leadingWhitespace := func(line string) string {
+		return line[:len(line)-len(strings.TrimLeft(line, " \t"))]
+	}
+	delta := ""
+	haveDelta := false
+	for i := range findLines {
+		spanLine := strings.TrimSuffix(spanLines[i], "\r")
+		findLine := strings.TrimSuffix(findLines[i], "\r")
+		if strings.TrimSpace(spanLine) == "" && strings.TrimSpace(findLine) == "" {
+			continue
+		}
+		spanIndent := leadingWhitespace(spanLine)
+		findIndent := leadingWhitespace(findLine)
+		if !strings.HasSuffix(spanIndent, findIndent) {
+			return "", false
+		}
+		lineDelta := spanIndent[:len(spanIndent)-len(findIndent)]
+		if !haveDelta {
+			delta = lineDelta
+			haveDelta = true
+			continue
+		}
+		if lineDelta != delta {
+			return "", false
+		}
+	}
+	return delta, haveDelta
+}
+
+// levenshtein computes edit distance with a two-row rolling matrix.
+func levenshtein(a, b string) int {
+	if a == "" {
+		return len(b)
+	}
+	if b == "" {
+		return len(a)
+	}
+	previous := make([]int, len(b)+1)
+	current := make([]int, len(b)+1)
+	for j := range previous {
+		previous[j] = j
+	}
+	for i := 1; i <= len(a); i++ {
+		current[0] = i
+		for j := 1; j <= len(b); j++ {
+			cost := 1
+			if a[i-1] == b[j-1] {
+				cost = 0
+			}
+			minimum := previous[j] + 1
+			if current[j-1]+1 < minimum {
+				minimum = current[j-1] + 1
+			}
+			if previous[j-1]+cost < minimum {
+				minimum = previous[j-1] + cost
+			}
+			current[j] = minimum
+		}
+		previous, current = current, previous
+	}
+	return previous[len(b)]
+}

internal/tools/edit_replacers_test.go

@@ -0,0 +1,369 @@
+package tools
+
+import (
+	"context"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+func runEdit(t *testing.T, dir, initial string, args map[string]any) (Result, string) {
+	t.Helper()
+	path := filepath.Join(dir, "target.txt")
+	if err := os.WriteFile(path, []byte(initial), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	if args["path"] == nil {
+		args["path"] = "target.txt"
+	}
+	result := NewEditFileTool(dir).Run(context.Background(), args)
+	after, err := os.ReadFile(path)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return result, string(after)
+}
+
+func TestEditExactMatchStillPreferred(t *testing.T) {
+	// The fast path must keep byte-exact semantics: an exact match wins even
+	// when fuzzy strategies would also find candidates elsewhere.
+	result, after := runEdit(t, t.TempDir(), "alpha\nbeta\ngamma\n", map[string]any{
+		"old_string": "beta",
+		"new_string": "delta",
+	})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %q", result.Output)
+	}
+	if after != "alpha\ndelta\ngamma\n" {
+		t.Fatalf("unexpected content: %q", after)
+	}
+}
+
+func TestEditFuzzyLineTrimmed(t *testing.T) {
+	// Model reproduced the lines without the file's leading indentation; the
+	// line-trimmed strategy must find the real indented span and preserve the
+	// file's own indentation on the replaced region boundary.
+	initial := "func a() {\n\tx := 1\n\treturn x\n}\n"
+	result, after := runEdit(t, t.TempDir(), initial, map[string]any{
+		"old_string": "x := 1\nreturn x",
+		"new_string": "y := 2\nreturn y",
+	})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %q", result.Output)
+	}
+	// new_string was written at old_string's outdented shape; the span's
+	// indentation must be re-applied so the file keeps its tabs.
+	if !strings.Contains(after, "\ty := 2\n\treturn y\n}") {
+		t.Fatalf("unexpected content: %q", after)
+	}
+	if strings.Contains(after, "x := 1") {
+		t.Fatalf("old span survived: %q", after)
+	}
+}
+
+func TestEditFuzzyWhitespaceNormalizedSingleLine(t *testing.T) {
+	// Whitespace runs collapsed: "x  :=   1" in the file, single spaces from
+	// the model.
+	initial := "start\nx  :=   1\nend\n"
+	result, after := runEdit(t, t.TempDir(), initial, map[string]any{
+		"old_string": "x := 1",
+		"new_string": "x := 2",
+	})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %q", result.Output)
+	}
+	if !strings.Contains(after, "x := 2") || strings.Contains(after, ":=   1") {
+		t.Fatalf("unexpected content: %q", after)
+	}
+}
+
+func TestEditFuzzyBlockAnchorToleratesMiddleDrift(t *testing.T) {
+	// First and last lines anchor the block; one interior line differs slightly
+	// (comment text drifted). Levenshtein similarity keeps it above 0.65.
+	initial := strings.Join([]string{
+		"func handler(w http.ResponseWriter, r *http.Request) {",
+		"\t// write the response body to the client",
+		"\tw.WriteHeader(http.StatusOK)",
+		"\tfmt.Fprint(w, \"done\")",
+		"}",
+		"",
+	}, "\n")
+	find := strings.Join([]string{
+		"func handler(w http.ResponseWriter, r *http.Request) {",
+		"\t// write the response body to client",
+		"\tw.WriteHeader(http.StatusOK)",
+		"\tfmt.Fprint(w, \"done\")",
+		"}",
+	}, "\n")
+	result, after := runEdit(t, t.TempDir(), initial, map[string]any{
+		"old_string": find,
+		"new_string": "func handler(w http.ResponseWriter, r *http.Request) {\n\tw.WriteHeader(http.StatusNoContent)\n}",
+	})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %q", result.Output)
+	}
+	if !strings.Contains(after, "StatusNoContent") || strings.Contains(after, "StatusOK") {
+		t.Fatalf("unexpected content: %q", after)
+	}
+}
+
+func TestEditFuzzyIndentationFlexible(t *testing.T) {
+	// The whole block sits one nesting level deeper in the file than in the
+	// model's old_string, with interior relative indentation preserved.
+	initial := "if ok {\n\t\tfor i := range xs {\n\t\t\tsum += xs[i]\n\t\t}\n}\n"
+	result, after := runEdit(t, t.TempDir(), initial, map[string]any{
+		"old_string": "for i := range xs {\n\tsum += xs[i]\n}",
+		"new_string": "sum += total(xs)",
+	})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %q", result.Output)
+	}
+	// The replacement lands at the span's real depth (two tabs), not at
+	// old_string's outdented depth.
+	if !strings.Contains(after, "\t\tsum += total(xs)") || strings.Contains(after, "range xs") {
+		t.Fatalf("unexpected content: %q", after)
+	}
+}
+
+func TestEditFuzzyCRLFLineTrimmedPreservesCarriageReturns(t *testing.T) {
+	// CRLF file + a multi-line outdented old_string: the resolved span carries
+	// tabs and trailing CRs; the replacement must inherit both so the file's
+	// CRLF pairs and indentation survive the edit.
+	initial := "func a() {\r\n\tx := 1\r\n\treturn x\r\n}\r\n"
+	result, after := runEdit(t, t.TempDir(), initial, map[string]any{
+		"old_string": "x := 1\nreturn x",
+		"new_string": "y := 2\nreturn y",
+	})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %q", result.Output)
+	}
+	if !strings.Contains(after, "\ty := 2\r\n\treturn y\r\n}") {
+		t.Fatalf("unexpected content: %q", after)
+	}
+	if strings.Contains(after, "\n\treturn y\n") {
+		t.Fatalf("bare LF introduced into CRLF file: %q", after)
+	}
+}
+
+func TestUniformIndentDelta(t *testing.T) {
+	cases := []struct {
+		span, find string
+		want       string
+		ok         bool
+	}{
+		{"\tx := 1\n\treturn x", "x := 1\nreturn x", "\t", true},
+		{"\t\tfor {\n\t\t\tgo()\n\t\t}", "for {\n\tgo()\n}", "\t\t", true},
+		{"x := 1", "x := 1", "", true},                          // no shift needed
+		{"\tx := 1\n  return x", "x := 1\nreturn x", "", false}, // mixed delta
+		{"\ta\n\tb\n\tc", "a\nb", "", false},                    // line-count mismatch
+		{"  a", "\ta", "", false},                               // find indent not a suffix
+	}
+	for _, c := range cases {
+		got, ok := uniformIndentDelta(c.span, c.find)
+		if ok != c.ok || (ok && got != c.want) {
+			t.Errorf("uniformIndentDelta(%q, %q) = %q,%v want %q,%v", c.span, c.find, got, ok, c.want, c.ok)
+		}
+	}
+}
+
+func TestAdaptReplacementLeavesNonUniformSpansAlone(t *testing.T) {
+	// Block-anchor style match with a drifted interior: no uniform delta, so
+	// the replacement must pass through untouched.
+	span := "\tfunc h() {\n\t\t// drifted comment\n\t}"
+	find := "func h() {\n// different comment\n}"
+	if got := adaptReplacementToSpan(span, find, "replacement()"); got != "replacement()" {
+		t.Fatalf("non-uniform span must not re-indent: %q", got)
+	}
+}
+
+func TestEditFuzzyEscapeNormalized(t *testing.T) {
+	// Model reproduced the line as it would appear inside a quoted string
+	// literal, with escaped quotes.
+	initial := "console.log(\"hello world\")\n"
+	result, after := runEdit(t, t.TempDir(), initial, map[string]any{
+		"old_string": `console.log(\"hello world\")`,
+		"new_string": `console.log("goodbye")`,
+	})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %q", result.Output)
+	}
+	if !strings.Contains(after, `console.log("goodbye")`) {
+		t.Fatalf("unexpected content: %q", after)
+	}
+}
+
+func TestEditFuzzyTrimmedBoundary(t *testing.T) {
+	// Stray whitespace around an otherwise-exact old_string; "  two  " is not a
+	// substring of the file, so only the fuzzy cascade can resolve it.
+	initial := "one\ntwo\nthree\n"
+	result, after := runEdit(t, t.TempDir(), initial, map[string]any{
+		"old_string": "  two  ",
+		"new_string": "2",
+	})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %q", result.Output)
+	}
+	if after != "one\n2\nthree\n" {
+		t.Fatalf("unexpected content: %q", after)
+	}
+}
+
+func TestTrimmedBoundaryReplacerYieldsTrimmedSpan(t *testing.T) {
+	candidates := trimmedBoundaryReplacer("one\ntwo\nthree\n", "\ntwo\n")
+	if len(candidates) == 0 || candidates[0] != "two" {
+		t.Fatalf("expected trimmed candidate \"two\", got %v", candidates)
+	}
+	if trimmedBoundaryReplacer("anything", "already-trimmed") != nil {
+		t.Fatal("already-trimmed find must yield no candidates")
+	}
+}
+
+// Two identical blocks that only match old_string after indentation-tolerant
+// matching: exact search finds nothing, fuzzy finds both.
+const ambiguousBlocks = "\tif x {\n\t\tgo()\n\t}\nmid\n\tif x {\n\t\tgo()\n\t}\n"
+
+func TestEditFuzzyAmbiguousReportsError(t *testing.T) {
+	// No strategy can disambiguate identical blocks, so the tool must refuse
+	// rather than guess, and the file must be untouched.
+	result, after := runEdit(t, t.TempDir(), ambiguousBlocks, map[string]any{
+		"old_string": "if x {\n\tgo()\n}",
+		"new_string": "stop()",
+	})
+	if result.Status != StatusError || !strings.Contains(result.Output, "multiple locations") {
+		t.Fatalf("expected ambiguity error, got %q", result.Output)
+	}
+	if after != ambiguousBlocks {
+		t.Fatalf("file must be unchanged, got %q", after)
+	}
+}
+
+func TestEditFuzzyReplaceAll(t *testing.T) {
+	// replace_all applies the fuzzy-resolved span at every occurrence.
+	result, after := runEdit(t, t.TempDir(), ambiguousBlocks, map[string]any{
+		"old_string":  "if x {\n\tgo()\n}",
+		"new_string":  "stop()",
+		"replace_all": true,
+	})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %q", result.Output)
+	}
+	if strings.Count(after, "stop()") != 2 || strings.Contains(after, "go()") {
+		t.Fatalf("unexpected content: %q", after)
+	}
+}
+
+func TestEditFuzzyNotFoundKeepsExactError(t *testing.T) {
+	result, after := runEdit(t, t.TempDir(), "alpha\n", map[string]any{
+		"old_string": "omega",
+		"new_string": "x",
+	})
+	if result.Status != StatusError || !strings.Contains(result.Output, "Could not find the exact string") {
+		t.Fatalf("expected not-found error, got %q", result.Output)
+	}
+	if after != "alpha\n" {
+		t.Fatalf("file must be unchanged, got %q", after)
+	}
+}
+
+func TestEditFuzzyCRLFFile(t *testing.T) {
+	// CRLF file + LF old_string with indentation drift: the CRLF translation
+	// feeds the cascade and the replacement preserves CRLF endings.
+	initial := "func a() {\r\n\tx := 1\r\n}\r\n"
+	result, after := runEdit(t, t.TempDir(), initial, map[string]any{
+		"old_string": "x := 1",
+		"new_string": "x := 2",
+	})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %q", result.Output)
+	}
+	if !strings.Contains(after, "\tx := 2\r\n") {
+		t.Fatalf("unexpected content: %q", after)
+	}
+}
+
+func TestIsDisproportionateEditMatch(t *testing.T) {
+	// A candidate span that dwarfs old_string must be refused: anchors bridging
+	// unrelated code would otherwise delete it all.
+	big := strings.Repeat("line\n", 10)
+	if !isDisproportionateEditMatch(big, "a\nb\nc") {
+		t.Fatal("10-line span for 3-line find must be disproportionate")
+	}
+	if isDisproportionateEditMatch("a\nb\nc\nd", "a\nb\nc") {
+		t.Fatal("one extra line within tolerance must be allowed")
+	}
+	if isDisproportionateEditMatch("single line", "single") {
+		t.Fatal("single-line finds are exempt from the byte-length guard")
+	}
+	if !isDisproportionateEditMatch("xx\n"+strings.Repeat("y", 900)+"\nzz", "xx\nab\nzz") {
+		t.Fatal("byte-length blowup must be disproportionate")
+	}
+}
+
+func TestLevenshtein(t *testing.T) {
+	cases := []struct {
+		a, b string
+		want int
+	}{
+		{"", "", 0},
+		{"abc", "", 3},
+		{"", "abc", 3},
+		{"kitten", "sitting", 3},
+		{"same", "same", 0},
+	}
+	for _, c := range cases {
+		if got := levenshtein(c.a, c.b); got != c.want {
+			t.Fatalf("levenshtein(%q,%q) = %d, want %d", c.a, c.b, got, c.want)
+		}
+	}
+}
+
+func TestEditFuzzyDistinctCandidatesAreAmbiguous(t *testing.T) {
+	// Two same-content blocks at DIFFERENT indentation: each resolved span is
+	// distinct and occurs exactly once, so the old literal-uniqueness check
+	// passed and silently edited the first block. Distinct-shaped candidates
+	// from one strategy must instead report ambiguity and leave the file alone.
+	initial := "\tif x {\n\t\tgo()\n\t}\nmid\n\t\tif x {\n\t\t\tgo()\n\t\t}\n"
+	result, after := runEdit(t, t.TempDir(), initial, map[string]any{
+		"old_string": "if x {\n\tgo()\n}",
+		"new_string": "stop()",
+	})
+	if result.Status != StatusError || !strings.Contains(result.Output, "multiple locations") {
+		t.Fatalf("distinct fuzzy candidates must be ambiguous, got %q", result.Output)
+	}
+	if after != initial {
+		t.Fatalf("file must be unchanged, got %q", after)
+	}
+}
+
+func TestEditFuzzyBlockAnchorTwoPlausibleBlocksAmbiguous(t *testing.T) {
+	// Two blocks share the same first/last anchor lines and BOTH interiors sit
+	// above the similarity threshold. Picking the "best" one would silently
+	// edit a block the model may not have meant; both must surface so the
+	// cascade reports ambiguity and the file stays untouched.
+	initial := strings.Join([]string{
+		"func setup(cfg Config) {",
+		"\tvalue := compute(alpha, beta, gamma)",
+		"}",
+		"",
+		"func setup(cfg Config) {",
+		"\tvalue := compute(alpha, delta, gamma)",
+		"}",
+		"",
+	}, "\n")
+	find := strings.Join([]string{
+		"func setup(cfg Config) {",
+		"\tvalue := compute(alpha, omega, gamma)",
+		"}",
+	}, "\n")
+	result, after := runEdit(t, t.TempDir(), initial, map[string]any{
+		"old_string": find,
+		"new_string": "func setup(cfg Config) {}",
+	})
+	if result.Status != StatusError || !strings.Contains(result.Output, "multiple locations") {
+		t.Fatalf("two plausible anchor blocks must be ambiguous, got %q", result.Output)
+	}
+	if after != initial {
+		t.Fatalf("file must be unchanged, got %q", after)
+	}
+}

internal/tools/exec_command.go

@@ -517,7 +517,12 @@ func (tool execCommandTool) run(ctx context.Context, args map[string]any, engine
 	if err != nil {
 		return errorResult("Error: Invalid arguments for exec_command: " + err.Error())
 	}
-	if issue := detectShellCommandIssue(commandText, runtimeGOOS()); issue != nil {
+	// Resolve the command engine before the MSYS preflight check so an
+	// approved require_escalated call (commandEngine == nil, truly
+	// unsandboxed) can actually bypass the MSYS guard instead of being
+	// hard-blocked by the same check it was meant to escalate past.
+	commandEngine := commandEngineForSandboxPermissions(engine, sandboxPermissions)
+	if issue := detectShellCommandIssue(commandText, runtimeGOOS()); issue != nil && !msysGuardBypassed(issue, commandEngine) {
 		return shellIssueBlockResult(*issue)
 	}
 	if interactive := zeroSandbox.DetectInteractiveCommand(commandText, runtimeGOOS()); interactive.Interactive {
@@ -860,6 +865,12 @@ func execToolResult(input execToolResultInput) Result {
 		status = StatusError
 	}
 	body := formatExecCommandOutput(output, input.sessionID, input.exited, input.exitCode, input.interrupted)
+	if status == StatusError && input.exited && !input.interrupted {
+		if issue := detectShellOutputIssue(output, runtimeGOOS()); issue != nil {
+			meta["shell_issue"] = issue.Kind
+			body = appendShellIssueHint(body, *issue)
+		}
+	}
 	if input.outputBufferTruncated {
 		// Appended after truncateExecOutput's own head/tail slicing, not
 		// embedded in the text that goes through it — a marker inside that
@@ -908,6 +919,15 @@ func formatExecCommandOutput(output string, sessionID int, exited bool, exitCode
 }
 
 func truncateExecOutput(output string, maxOutputTokens int) (string, bool) {
+	return truncateExecOutputSpill(output, maxOutputTokens, "exec_command")
+}
+
+// truncateExecOutputSpill keeps a head/tail window of the output within the
+// token budget and, on truncation, spills the full output to disk so the model
+// can grep/read the elided middle instead of re-running the command with a
+// bigger budget. The spill is best-effort: when it fails the notice simply
+// omits the file hint.
+func truncateExecOutputSpill(output string, maxOutputTokens int, toolName string) (string, bool) {
 	if maxOutputTokens <= 0 {
 		maxOutputTokens = defaultMaxOutputTokens
 	}
@@ -915,9 +935,13 @@ func truncateExecOutput(output string, maxOutputTokens int) (string, bool) {
 	if len(output) <= maxBytes {
 		return output, false
 	}
+	notice := "\n[zero] output truncated\n"
+	if spillPath := spillTruncatedOutput(toolName, output); spillPath != "" {
+		notice = "\n[zero] output truncated — full output saved to " + spillPath + " (grep or read_file it instead of re-running)\n"
+	}
 	head := maxBytes / 2
 	tail := maxBytes - head
-	return utf8Prefix(output, head) + "\n[zero] output truncated\n" + utf8Suffix(output, tail), true
+	return utf8Prefix(output, head) + notice + utf8Suffix(output, tail), true
 }
 
 func utf8Prefix(value string, maxBytes int) string {

internal/tools/exec_command_test.go

@@ -141,6 +141,46 @@ func TestExecCommandRequireEscalatedBypassesNativeSandboxAfterApproval(t *testin
 	}
 }
 
+// TestExecCommandRequireEscalatedBypassesMsysGuardAfterApproval mirrors
+// TestBashToolRequireEscalatedMsysGuard for exec_command: the MSYS sandbox
+// guard exists only because MSYS/Cygwin coreutils fail under the
+// write-restricted sandbox, so once require_escalated is actually approved
+// (unsandboxed execution), the guard must not block the same command it was
+// meant to let escalate past.
+func TestExecCommandRequireEscalatedBypassesMsysGuardAfterApproval(t *testing.T) {
+	if runtime.GOOS != "windows" {
+		t.Skip("windows-only MSYS sandbox guard")
+	}
+	root := t.TempDir()
+	manager := newExecSessionManager()
+	registry := NewRegistry()
+	registry.Register(NewScopedExecCommandTool(root, nil, manager))
+	engine := sandbox.NewEngine(sandbox.EngineOptions{
+		WorkspaceRoot: root,
+		Policy:        sandbox.DefaultPolicy(),
+		Backend:       sandbox.Backend{Name: sandbox.BackendUnavailable, Message: "native sandbox unavailable"},
+	})
+
+	result := registry.RunWithOptions(context.Background(), ExecCommandToolName, map[string]any{
+		"cmd":                 "cat somefile.txt",
+		"sandbox_permissions": string(SandboxPermissionsRequireEscalated),
+	}, RunOptions{
+		PermissionGranted: true,
+		Sandbox:           engine,
+		PermissionMode:    string(sandbox.PermissionModeAsk),
+	})
+
+	// Assert on the preflight block sentinel (exit_code "-1", set only by
+	// shellIssueBlockResult) rather than shell_issue: once the guard is
+	// bypassed, "cat somefile.txt" actually runs, and its real,
+	// PATH-dependent output could otherwise trip the unrelated
+	// post-execution detectShellOutputIssue heuristic and make this
+	// assertion flaky for reasons unrelated to the guard under test.
+	if result.Meta["exit_code"] == "-1" {
+		t.Fatalf("expected approved require_escalated to bypass the MSYS guard, got blocked: %#v", result)
+	}
+}
+
 func TestExecCommandReturnsExitCodeWhenCommandCompletesDuringInitialYield(t *testing.T) {
 	root := t.TempDir()
 	manager := newExecSessionManager()

internal/tools/format_on_write.go

@@ -0,0 +1,97 @@
+package tools
+
+import (
+	"context"
+	"os"
+	"os/exec"
+	"path/filepath"
+	"strings"
+	"time"
+)
+
+// Format-on-write for the mutating file tools. When enabled, a successful
+// edit_file/write_file runs the language's standard formatter on the file it
+// just wrote, so the model's output always lands in project-canonical style
+// and never fails a CI format check it cannot see. Off by default (set
+// ZERO_FORMAT_ON_WRITE=1): auto-reformatting changes bytes the model did not
+// write, which strict workflows may not want.
+//
+// Ordering matters: formatting runs BEFORE the FileTracker re-baseline, and
+// the caller records the POST-format content. Formatting after the baseline
+// would make the very next edit look like an external modification and trip
+// the conflict guard.
+
+// formatOnWriteTimeout bounds one formatter run; a wedged formatter must never
+// hang a tool call. On timeout the unformatted write stands.
+const formatOnWriteTimeout = 10 * time.Second
+
+// formatterCommands maps a file extension to the formatter argv; the file path
+// is appended as the final argument. Only in-place, config-respecting,
+// community-standard formatters — a missing binary silently skips formatting.
+var formatterCommands = map[string][]string{
+	".go":    {"gofmt", "-w"},
+	".rs":    {"rustfmt"},
+	".py":    {"ruff", "format", "--quiet"},
+	".ts":    {"prettier", "--log-level", "silent", "--write"},
+	".tsx":   {"prettier", "--log-level", "silent", "--write"},
+	".js":    {"prettier", "--log-level", "silent", "--write"},
+	".jsx":   {"prettier", "--log-level", "silent", "--write"},
+	".json":  {"prettier", "--log-level", "silent", "--write"},
+	".css":   {"prettier", "--log-level", "silent", "--write"},
+	".scss":  {"prettier", "--log-level", "silent", "--write"},
+	".html":  {"prettier", "--log-level", "silent", "--write"},
+	".md":    {"prettier", "--log-level", "silent", "--write"},
+	".yaml":  {"prettier", "--log-level", "silent", "--write"},
+	".yml":   {"prettier", "--log-level", "silent", "--write"},
+	".zig":   {"zig", "fmt"},
+	".dart":  {"dart", "format"},
+	".tf":    {"terraform", "fmt"},
+	".gleam": {"gleam", "format"},
+	".sh":    {"shfmt", "-w"},
+	".bash":  {"shfmt", "-w"},
+	".c":     {"clang-format", "-i"},
+	".h":     {"clang-format", "-i"},
+	".cpp":   {"clang-format", "-i"},
+	".hpp":   {"clang-format", "-i"},
+	".cc":    {"clang-format", "-i"},
+	".kt":    {"ktlint", "-F"},
+	".swift": {"swiftformat"},
+	".lua":   {"stylua"},
+}
+
+// formatOnWriteEnabled reports whether the opt-in env toggle is set.
+func formatOnWriteEnabled() bool {
+	value := strings.TrimSpace(os.Getenv("ZERO_FORMAT_ON_WRITE"))
+	return value != "" && value != "0" && !strings.EqualFold(value, "false")
+}
+
+// maybeFormatWrittenFile runs the configured formatter for absolutePath (when
+// enabled and on PATH) and returns the file's content afterwards. Best-effort
+// throughout: any failure — no formatter, formatter error, timeout, unreadable
+// result — returns writtenContent so the caller's state matches the last write
+// it performed itself.
+func maybeFormatWrittenFile(ctx context.Context, absolutePath string, writtenContent string) string {
+	if !formatOnWriteEnabled() {
+		return writtenContent
+	}
+	command, ok := formatterCommands[strings.ToLower(filepath.Ext(absolutePath))]
+	if !ok {
+		return writtenContent
+	}
+	if _, err := exec.LookPath(command[0]); err != nil {
+		return writtenContent
+	}
+	formatCtx, cancel := context.WithTimeout(ctx, formatOnWriteTimeout)
+	defer cancel()
+	arguments := append(append([]string(nil), command[1:]...), absolutePath)
+	formatter := exec.CommandContext(formatCtx, command[0], arguments...)
+	formatter.Dir = filepath.Dir(absolutePath)
+	if err := formatter.Run(); err != nil {
+		return writtenContent
+	}
+	formatted, err := os.ReadFile(absolutePath)
+	if err != nil {
+		return writtenContent
+	}
+	return string(formatted)
+}

internal/tools/format_on_write_test.go

@@ -0,0 +1,82 @@
+package tools
+
+import (
+	"context"
+	"os"
+	"os/exec"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+// gofmt ships with the Go toolchain, so it is the one formatter guaranteed to
+// exist wherever these tests run.
+func requireGofmt(t *testing.T) {
+	t.Helper()
+	if _, err := exec.LookPath("gofmt"); err != nil {
+		t.Skip("gofmt not on PATH")
+	}
+}
+
+func TestFormatOnWriteDisabledByDefault(t *testing.T) {
+	requireGofmt(t)
+	t.Setenv("ZERO_FORMAT_ON_WRITE", "")
+	dir := t.TempDir()
+	ugly := "package a\n\nfunc  A( ) {   }\n"
+
+	result := NewWriteFileTool(dir).(optionsAwareTool).RunWithOptions(context.Background(), map[string]any{
+		"path":    "a.go",
+		"content": ugly,
+	}, RunOptions{})
+	if result.Status != StatusOK {
+		t.Fatalf("write failed: %q", result.Output)
+	}
+	onDisk, err := os.ReadFile(filepath.Join(dir, "a.go"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if string(onDisk) != ugly {
+		t.Fatalf("formatting must be off by default, got %q", onDisk)
+	}
+}
+
+func TestFormatOnWriteFormatsAndKeepsTrackerConsistent(t *testing.T) {
+	requireGofmt(t)
+	t.Setenv("ZERO_FORMAT_ON_WRITE", "1")
+	dir := t.TempDir()
+	tracker := NewFileTracker()
+
+	write := NewWriteFileTool(dir).(optionsAwareTool).RunWithOptions(context.Background(), map[string]any{
+		"path":    "a.go",
+		"content": "package a\n\nfunc  A( ) {   }\n",
+	}, RunOptions{FileTracker: tracker})
+	if write.Status != StatusOK {
+		t.Fatalf("write failed: %q", write.Output)
+	}
+	onDisk, err := os.ReadFile(filepath.Join(dir, "a.go"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if !strings.Contains(string(onDisk), "func A() {") {
+		t.Fatalf("expected gofmt-formatted content, got %q", onDisk)
+	}
+
+	// The tracker must have been re-baselined to the POST-format content: a
+	// follow-up edit must not trip the external-modification conflict guard.
+	edit := NewEditFileTool(dir).(optionsAwareTool).RunWithOptions(context.Background(), map[string]any{
+		"path":       "a.go",
+		"old_string": "func A() {",
+		"new_string": "func B() {",
+	}, RunOptions{FileTracker: tracker})
+	if edit.Status != StatusOK {
+		t.Fatalf("follow-up edit must not conflict after formatting: %q", edit.Output)
+	}
+}
+
+func TestFormatOnWriteSkipsUnknownExtensions(t *testing.T) {
+	t.Setenv("ZERO_FORMAT_ON_WRITE", "1")
+	content := maybeFormatWrittenFile(context.Background(), filepath.Join(t.TempDir(), "notes.xyz"), "raw   text")
+	if content != "raw   text" {
+		t.Fatalf("unknown extension must pass through: %q", content)
+	}
+}

internal/tools/inline_diagnostics.go

@@ -0,0 +1,18 @@
+package tools
+
+import "context"
+
+// inlineDiagnostics renders the post-write diagnostics block a mutating tool
+// appends to its output. Empty when no Diagnostics callback is wired, the file
+// is clean, or no language server is available — the tool output is then
+// byte-identical to the pre-diagnostics behavior.
+func inlineDiagnostics(ctx context.Context, options RunOptions, absolutePath, relativePath string) string {
+	if options.Diagnostics == nil {
+		return ""
+	}
+	block := options.Diagnostics(ctx, absolutePath)
+	if block == "" {
+		return ""
+	}
+	return "\n\nDiagnostics in " + relativePath + " after this change (fix any errors you introduced):\n" + block
+}

internal/tools/inline_diagnostics_test.go

@@ -0,0 +1,53 @@
+package tools
+
+import (
+	"context"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+// A wired Diagnostics callback must surface its block in the tool output for
+// both mutating tools, and a clean file (empty block) must leave the output
+// byte-identical to the pre-diagnostics behavior.
+func TestMutatingToolsAppendInlineDiagnostics(t *testing.T) {
+	dir := t.TempDir()
+	if err := os.WriteFile(filepath.Join(dir, "a.go"), []byte("package a\n"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	diagnostics := func(_ context.Context, absPath string) string {
+		return absPath + ":1:1: error: undefined: x"
+	}
+
+	editResult := NewEditFileTool(dir).(optionsAwareTool).RunWithOptions(context.Background(), map[string]any{
+		"path":       "a.go",
+		"old_string": "package a",
+		"new_string": "package b",
+	}, RunOptions{Diagnostics: diagnostics})
+	if editResult.Status != StatusOK {
+		t.Fatalf("edit failed: %q", editResult.Output)
+	}
+	if !strings.Contains(editResult.Output, "Diagnostics in a.go after this change") || !strings.Contains(editResult.Output, "undefined: x") {
+		t.Fatalf("edit output missing diagnostics block: %q", editResult.Output)
+	}
+
+	writeResult := NewWriteFileTool(dir).(optionsAwareTool).RunWithOptions(context.Background(), map[string]any{
+		"path":    "b.go",
+		"content": "package b\n",
+	}, RunOptions{Diagnostics: diagnostics})
+	if writeResult.Status != StatusOK {
+		t.Fatalf("write failed: %q", writeResult.Output)
+	}
+	if !strings.Contains(writeResult.Output, "Diagnostics in b.go after this change") {
+		t.Fatalf("write output missing diagnostics block: %q", writeResult.Output)
+	}
+
+	clean := NewWriteFileTool(dir).(optionsAwareTool).RunWithOptions(context.Background(), map[string]any{
+		"path":    "c.go",
+		"content": "package c\n",
+	}, RunOptions{Diagnostics: func(context.Context, string) string { return "" }})
+	if strings.Contains(clean.Output, "Diagnostics") {
+		t.Fatalf("clean file must not gain a diagnostics block: %q", clean.Output)
+	}
+}

internal/tools/output_ceiling.go

@@ -0,0 +1,81 @@
+package tools
+
+import (
+	"os"
+	"strconv"
+	"strings"
+)
+
+// Universal output ceiling. Every tool result leaving the registry boundary is
+// capped at a single token-denominated budget unless the tool manages its own
+// deliberate budget (selfBudgeting below). This is the safety net for tools
+// with no cap of their own — web_fetch bodies, skill files, MCP server tools,
+// browser snapshots, anything added later — so no single call can flood the
+// context window with output that is then re-billed on every following turn.
+// Truncation keeps head+tail and spills the full text to disk (re-readable via
+// grep/read_file), so nothing is lost, only deferred.
+
+// defaultOutputCeilingTokens is the per-call ceiling in estimated tokens
+// (bytes/4). 16k tokens = 64 KiB — deliberately equal to the search-tool
+// budget: generous enough for a large fetched document, small enough that one
+// call cannot eat a third of a small context window.
+const defaultOutputCeilingTokens = 16_000
+
+// outputCeilingEnv overrides the ceiling (in tokens). Zero or negative
+// disables the ceiling entirely; unset or unparsable keeps the default.
+const outputCeilingEnv = "ZERO_TOOL_OUTPUT_CEILING_TOKENS"
+
+// selfBudgeting marks a tool that enforces its own deliberate output budget —
+// possibly model-raisable (exec_command) — which the registry ceiling must not
+// second-guess. The method is unexported on purpose: only tools in this
+// package can opt out, so an MCP-served tool can never exempt itself.
+type selfBudgeting interface{ managesOutputBudget() }
+
+// The exemption list, kept in one place. Each of these applies its own budget
+// before returning: bash (bashOutputBudgetBytes per stream + spill),
+// exec_command (model-raisable token budget + spill), read tools (128 KiB),
+// search tools (64 KiB).
+func (bashTool) managesOutputBudget()             {}
+func (execCommandTool) managesOutputBudget()      {}
+func (readFileTool) managesOutputBudget()         {}
+func (readMinifiedFileTool) managesOutputBudget() {}
+func (grepTool) managesOutputBudget()             {}
+func (globTool) managesOutputBudget()             {}
+func (listDirectoryTool) managesOutputBudget()    {}
+
+func resolveOutputCeilingTokens() int {
+	raw := strings.TrimSpace(os.Getenv(outputCeilingEnv))
+	if raw == "" {
+		return defaultOutputCeilingTokens
+	}
+	parsed, err := strconv.Atoi(raw)
+	if err != nil {
+		return defaultOutputCeilingTokens
+	}
+	return parsed
+}
+
+// enforceOutputCeiling truncates an over-ceiling result to head+tail within
+// the token budget, spilling the full (already redaction-scrubbed) output to
+// disk with a recovery hint. Runs after scrubResultSecrets so the transcript
+// and the spill file agree on what was hidden.
+func enforceOutputCeiling(toolName string, result Result) Result {
+	ceiling := resolveOutputCeilingTokens()
+	if ceiling <= 0 {
+		return result
+	}
+	if len(result.Output) <= ceiling*4 {
+		return result
+	}
+	rawBytes := len(result.Output)
+	truncated, _ := truncateExecOutputSpill(result.Output, ceiling, toolName)
+	result.Output = truncated
+	result.Truncated = true
+	if result.Meta == nil {
+		result.Meta = map[string]string{}
+	}
+	result.Meta["raw_bytes"] = strconv.Itoa(rawBytes)
+	result.Meta["emitted_bytes"] = strconv.Itoa(len(result.Output))
+	result.Meta["estimated_tokens"] = strconv.Itoa(estimatedTokensFromBytes(len(result.Output)))
+	return result
+}

internal/tools/output_ceiling_test.go

@@ -0,0 +1,140 @@
+package tools
+
+import (
+	"context"
+	"os"
+	"strconv"
+	"strings"
+	"testing"
+)
+
+// ceilingFakeTool is a minimal tool with no output budget of its own — the
+// stand-in for an MCP-served tool or any future builtin that forgets to cap
+// its output.
+type ceilingFakeTool struct {
+	baseTool
+	output string
+}
+
+func (tool ceilingFakeTool) Run(context.Context, map[string]any) Result {
+	return okResult(tool.output)
+}
+
+// ceilingExemptTool is the same fake but marked self-budgeting.
+type ceilingExemptTool struct{ ceilingFakeTool }
+
+func (ceilingExemptTool) managesOutputBudget() {}
+
+func newCeilingFakeTool(name, output string) ceilingFakeTool {
+	return ceilingFakeTool{
+		baseTool: baseTool{name: name, safety: readOnlySafety("test tool")},
+		output:   output,
+	}
+}
+
+// An unbudgeted tool's oversized output is capped at the universal ceiling,
+// head and tail survive, and the full output is spilled to a re-readable file.
+func TestOutputCeilingCapsUnbudgetedTool(t *testing.T) {
+	setTestTempDir(t)
+	big := "HEAD_MARK\n" + strings.Repeat("z", defaultOutputCeilingTokens*4*3) + "\nTAIL_MARK"
+
+	registry := NewRegistry()
+	registry.Register(newCeilingFakeTool("fake_big", big))
+	result := registry.Run(context.Background(), "fake_big", map[string]any{})
+
+	if !result.Truncated {
+		t.Fatal("over-ceiling output must report truncated=true")
+	}
+	maxBytes := defaultOutputCeilingTokens * 4
+	if len(result.Output) > maxBytes+512 {
+		t.Fatalf("emitted %d bytes, ceiling is %d", len(result.Output), maxBytes)
+	}
+	if !strings.Contains(result.Output, "HEAD_MARK") || !strings.Contains(result.Output, "TAIL_MARK") {
+		t.Fatal("head/tail lost at the ceiling")
+	}
+	if result.Meta["raw_bytes"] != strconv.Itoa(len(big)) {
+		t.Fatalf("raw_bytes = %s, want %d", result.Meta["raw_bytes"], len(big))
+	}
+	if !strings.Contains(result.Output, "full output saved to ") {
+		t.Fatalf("ceiling truncation must include a spill hint: %q", result.Output[:200])
+	}
+	start := strings.Index(result.Output, "full output saved to ") + len("full output saved to ")
+	end := strings.Index(result.Output[start:], " (grep")
+	content, err := os.ReadFile(result.Output[start : start+end])
+	if err != nil {
+		t.Fatalf("spill file unreadable: %v", err)
+	}
+	if string(content) != big {
+		t.Fatalf("spill must hold the full output: got %d bytes, want %d", len(content), len(big))
+	}
+}
+
+// A self-budgeting tool's output passes the boundary untouched even when it
+// exceeds the universal ceiling — its own (possibly model-raised) budget wins.
+func TestOutputCeilingSkipsSelfBudgetingTool(t *testing.T) {
+	big := strings.Repeat("z", defaultOutputCeilingTokens*4*2)
+	registry := NewRegistry()
+	registry.Register(ceilingExemptTool{newCeilingFakeTool("fake_exempt", big)})
+
+	result := registry.Run(context.Background(), "fake_exempt", map[string]any{})
+	if result.Truncated || result.Output != big {
+		t.Fatalf("self-budgeting tool must bypass the ceiling (truncated=%v, %d bytes)", result.Truncated, len(result.Output))
+	}
+}
+
+// Under-ceiling output is untouched and gains no budget metadata.
+func TestOutputCeilingSmallOutputUntouched(t *testing.T) {
+	registry := NewRegistry()
+	registry.Register(newCeilingFakeTool("fake_small", "hello"))
+
+	result := registry.Run(context.Background(), "fake_small", map[string]any{})
+	if result.Truncated || result.Output != "hello" {
+		t.Fatalf("small output altered: %+v", result)
+	}
+	if _, ok := result.Meta["raw_bytes"]; ok {
+		t.Fatal("small output must not gain budget metadata")
+	}
+}
+
+// ZERO_TOOL_OUTPUT_CEILING_TOKENS tightens, loosens, or disables the ceiling.
+func TestOutputCeilingEnvOverride(t *testing.T) {
+	setTestTempDir(t)
+	big := strings.Repeat("z", 100*1024)
+
+	t.Setenv(outputCeilingEnv, "1000")
+	registry := NewRegistry()
+	registry.Register(newCeilingFakeTool("fake_env", big))
+	result := registry.Run(context.Background(), "fake_env", map[string]any{})
+	if !result.Truncated || len(result.Output) > 1000*4+512 {
+		t.Fatalf("ceiling override ignored: truncated=%v, %d bytes", result.Truncated, len(result.Output))
+	}
+
+	t.Setenv(outputCeilingEnv, "0")
+	result = registry.Run(context.Background(), "fake_env", map[string]any{})
+	if result.Truncated || len(result.Output) != len(big) {
+		t.Fatalf("ceiling=0 must disable the net: truncated=%v, %d bytes", result.Truncated, len(result.Output))
+	}
+}
+
+// The exemption list matches the tools that really do budget themselves, and
+// the unbudgeted ones (web_fetch, skill, browser) stay under the net.
+func TestSelfBudgetingExemptionList(t *testing.T) {
+	dir := t.TempDir()
+	exempt := []Tool{
+		NewBashTool(dir),
+		NewExecCommandTool(dir, newExecSessionManager()),
+		NewReadFileTool(dir),
+		NewReadMinifiedFileTool(dir),
+		NewGrepTool(dir),
+		NewGlobTool(dir),
+		NewListDirectoryTool(dir),
+	}
+	for _, tool := range exempt {
+		if _, ok := tool.(selfBudgeting); !ok {
+			t.Errorf("%s must be exempt from the output ceiling", tool.Name())
+		}
+	}
+	if _, ok := NewWebFetchTool().(selfBudgeting); ok {
+		t.Error("web_fetch must NOT be exempt — the ceiling is its only budget")
+	}
+}

internal/tools/registry.go

@@ -40,6 +40,12 @@ type RunOptions struct {
 	// tool-call progress in the specialist card. nil is a no-op (the default
 	// for every non-Task tool).
 	Progress func(streamjson.Event)
+	// Diagnostics, when set, returns a formatted language-diagnostics block for
+	// a file a mutating tool just wrote ("" when clean or no server available).
+	// edit_file/write_file append it to their output so the model sees an error
+	// it introduced in the same turn instead of waiting for a later verification
+	// pass. nil disables inline diagnostics.
+	Diagnostics func(ctx context.Context, absPath string) string
 }
 
 type sandboxAwareTool interface {
@@ -117,13 +123,24 @@ func (registry *Registry) Run(ctx context.Context, name string, args map[string]
 func (registry *Registry) RunWithOptions(ctx context.Context, name string, args map[string]any, options RunOptions) (result Result) {
 	// Every return path passes through scrubResultSecrets exactly once, so denial,
 	// permission, and unknown-tool error messages (which can echo secret-bearing
-	// args/paths) are redacted at the boundary just like tool output.
-	defer func() { result = scrubResultSecrets(result) }()
+	// args/paths) are redacted at the boundary just like tool output. The output
+	// ceiling runs after the scrub so the transcript and the spill file agree on
+	// what was hidden.
+	ceilingExempt := false
+	defer func() {
+		result = scrubResultSecrets(result)
+		if !ceilingExempt {
+			result = enforceOutputCeiling(name, result)
+		}
+	}()
 
 	tool, ok := registry.Get(name)
 	if !ok {
 		return errorResult(`Error: Unknown tool "` + name + `".`)
 	}
+	if _, ok := tool.(selfBudgeting); ok {
+		ceilingExempt = true
+	}
 	if rejecter, ok := tool.(PrePermissionRejecter); ok {
 		if res, rejected := rejecter.RejectBeforePermission(args); rejected {
 			return res

internal/tools/shell_runtime.go

@@ -3,6 +3,7 @@ package tools
 import (
 	"regexp"
 	"strings"
+	"unicode"
 )
 
 type shellRuntime struct {
@@ -17,14 +18,31 @@ type shellIssue struct {
 	Suggestion string
 }
 
+const windowsMsysSandboxKind = "windows_msys_sandbox"
+
+const windowsMsysSandboxSuggestion = "MSYS/Cygwin coreutils from Git for Windows cannot run under Zero's write-restricted Windows sandbox. Prefer Zero native tools (grep, read_file with offset/limit, list_directory, glob), cmd.exe findstr/more, or PowerShell Select-Object -First/-Last. If host-level execution is truly required, rerun with sandbox_permissions: \"require_escalated\" and a narrow justification."
+
+// windowsMsysProneNames is the single source of truth for POSIX coreutil names
+// that commonly resolve to a Git-for-Windows MSYS/Cygwin binary rather than a
+// cmd.exe-native command, and so fail under the write-restricted Windows
+// sandbox (#458). Every Windows MSYS-detection path (the preflight command
+// scan below, the exported MsysProneCommandName, and the known-safe-segment
+// guard in internal/agent/command_prefix.go) derives from this one set, so
+// they cannot drift out of sync with each other.
+var windowsMsysProneNames = map[string]bool{
+	"cat": true, "cut": true, "expr": true, "grep": true, "head": true,
+	"id": true, "ls": true, "nl": true, "paste": true, "rev": true,
+	"seq": true, "stat": true, "tail": true, "tr": true, "uname": true,
+	"uniq": true, "wc": true, "which": true, "awk": true, "sed": true,
+	"xargs": true,
+}
+
 var (
 	windowsBashStyleCDPattern = regexp.MustCompile(`(?i)(^|[&|;]\s*)cd\s+/(?:[a-ce-z0-9_./~-]|d[a-z0-9_./~-])[a-z0-9_./~-]*`)
-	windowsLSCommandPattern   = regexp.MustCompile(`(?i)(^|[&|;]\s*)ls\b(?:\s+|$)`)
-	// windowsPosixUtilityPattern catches the other common POSIX coreutils/grep
-	// family commands cmd.exe has no equivalent for (unlike `dir`/`findstr`,
-	// which sound similar enough that a model might reach for the POSIX name
-	// instead) — most often seen piped in, e.g. `git log ... | head`.
-	windowsPosixUtilityPattern = regexp.MustCompile(`(?i)(^|[&|;]\s*)(head|tail|grep|wc|awk|sed|cut|xargs|tr)(?:\s+|$)`)
+	// windowsMsysBinaryPathPattern catches explicit Git-for-Windows / MSYS usr\bin
+	// paths. These executables are valid Windows PE files but fail under the
+	// write-restricted sandbox with CreateFileMapping ACCESS_DENIED (#458).
+	windowsMsysBinaryPathPattern = regexp.MustCompile(`(?i)(?:\\usr\\bin\\|\\mingw64\\bin\\|msys-2\.0\.dll|cygwin1\.dll)`)
 )
 
 func detectShellRuntime(goos string) shellRuntime {
@@ -37,48 +55,162 @@ func detectShellRuntime(goos string) shellRuntime {
 func shellGuidanceForGOOS(goos string) string {
 	runtime := detectShellRuntime(goos)
 	if goos == "windows" {
-		return "Uses " + runtime.Syntax + " syntax on Windows; prefer cwd over cd when changing directories."
+		return "Uses " + runtime.Syntax + " syntax on Windows; prefer cwd over cd when changing directories. MSYS/Cygwin coreutils on PATH (Git for Windows usr\\bin) are not sandbox-compatible; prefer native Zero file tools."
 	}
 	guidance := "Uses " + runtime.Syntax + " syntax."
 	if goos == "darwin" {
-		// `ps` is setuid root and cannot run under the macOS sandbox; `pgrep` needs a
-		// blocked system service. Point the model at the tools that DO work so it
-		// doesn't waste turns: lsof to find a process, kill to stop it.
 		guidance += " To find or stop a process, use `lsof -i :PORT` (or `lsof -nP -iTCP -sTCP:LISTEN`) for the PID then `kill <pid>`; `ps` and `pgrep` do not work under the sandbox."
 	}
 	return guidance
 }
 
+// MsysProneCommandName reports whether a bare command name commonly resolves to
+// a Git-for-Windows MSYS binary that fails under the Windows restricted sandbox.
+func MsysProneCommandName(name string) bool {
+	return windowsMsysProneNames[strings.ToLower(strings.TrimSpace(name))]
+}
+
+func windowsMsysSandboxIssue(message string) *shellIssue {
+	return &shellIssue{
+		Kind:       windowsMsysSandboxKind,
+		Message:    message,
+		Suggestion: windowsMsysSandboxSuggestion,
+	}
+}
+
+// windowsCommandSegments splits a command into cmd.exe-operator-separated
+// segments (&, |, and their doubled forms &&/||), respecting double quotes
+// (cmd.exe's grouping construct) and the caret (^) escape character, so an
+// operator or command name mentioned inside a quoted argument (e.g. a commit
+// message or PR comment body), or an operator escaped with ^ (cmd.exe prints
+// `echo ^| head` as literal text instead of piping to head), is not mistaken
+// for a real segment boundary or invocation. Unlike bash, cmd.exe does not
+// treat ; as a statement separator, so it is left as ordinary argument text
+// (e.g. `echo foo; head` is a single `echo` invocation with literal args).
+func windowsCommandSegments(command string) []string {
+	var segments []string
+	var current strings.Builder
+	inQuotes := false
+	runes := []rune(command)
+	for i := 0; i < len(runes); i++ {
+		c := runes[i]
+		if !inQuotes && c == '^' && i+1 < len(runes) {
+			current.WriteRune(c)
+			i++
+			current.WriteRune(runes[i])
+			continue
+		}
+		if c == '"' {
+			inQuotes = !inQuotes
+			current.WriteRune(c)
+			continue
+		}
+		if !inQuotes && (c == '&' || c == '|') {
+			if seg := strings.TrimSpace(current.String()); seg != "" {
+				segments = append(segments, seg)
+			}
+			current.Reset()
+			continue
+		}
+		current.WriteRune(c)
+	}
+	if seg := strings.TrimSpace(current.String()); seg != "" {
+		segments = append(segments, seg)
+	}
+	return segments
+}
+
+// firstCommandWord returns the first token of a command segment. A leading
+// double-quoted span counts as one token with its quotes stripped, since
+// cmd.exe treats a quoted path as a single argument: the command invoked by
+// `"C:\Program Files\Git\usr\bin\head.exe" file.txt` is the quoted path, not
+// "C:\Program. For an unquoted command, the token ends at whitespace or a
+// redirection operator (<, >): cmd.exe accepts redirection attached directly
+// to the command name with no separating space (head>out.txt, cat<in.txt), so
+// stopping only at whitespace would return "head>out.txt" as one word and
+// miss the invoked command.
+func firstCommandWord(segment string) string {
+	trimmed := strings.TrimSpace(segment)
+	if trimmed == "" {
+		return ""
+	}
+	if trimmed[0] == '"' {
+		if end := strings.IndexByte(trimmed[1:], '"'); end >= 0 {
+			return trimmed[1 : end+1]
+		}
+		return trimmed[1:]
+	}
+	if end := strings.IndexFunc(trimmed, isCommandWordBoundary); end >= 0 {
+		return trimmed[:end]
+	}
+	return trimmed
+}
+
+func isCommandWordBoundary(r rune) bool {
+	return unicode.IsSpace(r) || r == '<' || r == '>'
+}
+
+// msysProneCommandWord reports whether word (the first token of a command
+// segment, as returned by firstCommandWord) names an MSYS-prone coreutil,
+// bare or with a directory prefix and/or .exe suffix (head, head.exe,
+// C:\...\usr\bin\head.exe, ...).
+func msysProneCommandWord(word string) bool {
+	word = strings.Trim(word, `"`)
+	if i := strings.LastIndexAny(word, `\/`); i >= 0 {
+		word = word[i+1:]
+	}
+	word = strings.TrimSuffix(strings.ToLower(word), ".exe")
+	return MsysProneCommandName(word)
+}
+
 func detectShellCommandIssue(command string, goos string) *shellIssue {
 	if goos != "windows" {
 		return nil
 	}
 	trimmed := strings.TrimSpace(command)
-	if windowsBashStyleCDPattern.MatchString(trimmed) ||
-		windowsLSCommandPattern.MatchString(trimmed) {
+	if windowsBashStyleCDPattern.MatchString(trimmed) {
 		return &shellIssue{
 			Kind:       "windows_shell_syntax",
 			Message:    "Command looks like POSIX/Bash syntax, but Zero runs bash tool commands through Windows cmd.exe on this host.",
 			Suggestion: "Use the cwd argument instead of cd, use Windows cmd.exe syntax, or use native tools such as list_directory, read_file, grep, and glob.",
 		}
 	}
-	if windowsPosixUtilityPattern.MatchString(trimmed) {
-		return &shellIssue{
-			Kind:       "windows_shell_syntax",
-			Message:    "Command uses a POSIX utility (head/tail/grep/wc/awk/sed/cut/xargs/tr) that Windows cmd.exe does not have.",
-			Suggestion: "Use cmd.exe equivalents (e.g. `findstr` for grep, `more` to page output) or Zero's native tools (grep, read_file with offset/limit) instead of piping to a POSIX utility.",
+	// Check the first word of each operator-separated segment (not the raw
+	// text anywhere in the command) against the MSYS binary-path pattern and
+	// the single MSYS-prone name set, covering bare names (head), .exe names
+	// (head.exe), and directory-prefixed forms (C:\...\head.exe) uniformly.
+	// Being segment/word anchored rather than a whole-string regex or scan,
+	// neither check matches text that only appears inside a quoted argument
+	// (e.g. a commit message or PR comment body discussing head.exe).
+	for _, segment := range windowsCommandSegments(trimmed) {
+		word := firstCommandWord(segment)
+		if windowsMsysBinaryPathPattern.MatchString(word) {
+			return windowsMsysSandboxIssue("Command invokes an MSYS/Cygwin binary path that cannot run under Zero's Windows sandbox.")
+		}
+		if msysProneCommandWord(word) {
+			return windowsMsysSandboxIssue("Command uses a POSIX coreutil (head/tail/grep/cat/...) that commonly resolves to Git-for-Windows MSYS binaries incompatible with the Windows sandbox.")
 		}
 	}
 	return nil
 }
 
-func detectShellOutputIssue(command string, output string, goos string) *shellIssue {
+// detectShellOutputIssue looks for MSYS runtime crash markers and cmd.exe
+// syntax-error text in output only, never in the command that was run. The
+// command line is attacker/user-controlled argument text (e.g. a `gh pr
+// comment --body` quoting a sample failure), not something the shell
+// produced, so treating it as evidence would reintroduce the same
+// quoted-text false positives the preflight command-position check exists to
+// avoid, just after execution instead of before it.
+func detectShellOutputIssue(output string, goos string) *shellIssue {
 	if goos != "windows" {
 		return nil
 	}
-	haystack := strings.ToLower(command + "\n" + output)
-	if strings.Contains(haystack, "the syntax of the command is incorrect") ||
-		strings.Contains(haystack, "is not recognized as an internal or external command") {
+	lower := strings.ToLower(output)
+	if msysRuntimeFailedInOutput(lower) {
+		return windowsMsysSandboxIssue("An MSYS/Cygwin runtime failed under Zero's Windows sandbox (ACCESS_DENIED during MSYS startup).")
+	}
+	if strings.Contains(lower, "the syntax of the command is incorrect") ||
+		strings.Contains(lower, "is not recognized as an internal or external command") {
 		return &shellIssue{
 			Kind:       "windows_shell_syntax",
 			Message:    "Windows cmd.exe rejected the command syntax.",
@@ -88,6 +220,32 @@ func detectShellOutputIssue(command string, output string, goos string) *shellIs
 	return nil
 }
 
+func msysRuntimeFailedInOutput(lower string) bool {
+	if strings.Contains(lower, "fatal error - createfilemapping") {
+		return true
+	}
+	if strings.Contains(lower, "couldn't create signal pipe") && strings.Contains(lower, "win32 error 5") {
+		return true
+	}
+	if strings.Contains(lower, "cygheap_user::init") && strings.Contains(lower, "fatal error") {
+		return true
+	}
+	if strings.Contains(lower, "usr\\bin\\") && strings.Contains(lower, "fatal error") {
+		return true
+	}
+	if !strings.Contains(lower, "win32 error 5") || !strings.Contains(lower, "terminating") {
+		return false
+	}
+	// Anchor the broad win32-error-5 fallback to an MSYS-specific marker so
+	// unrelated access-denied failures are not mislabeled as MSYS sandbox
+	// incompatibilities.
+	return strings.Contains(lower, `usr\bin\`) ||
+		strings.Contains(lower, "cygheap") ||
+		strings.Contains(lower, "msys-2.0.dll") ||
+		strings.Contains(lower, "cygwin1.dll") ||
+		strings.Contains(lower, "[main]")
+}
+
 func appendShellIssueHint(output string, issue shellIssue) string {
 	output = strings.TrimRight(output, "\r\n")
 	hint := "[zero] shell issue: " + issue.Message

internal/tools/shell_runtime_test.go

@@ -0,0 +1,208 @@
+package tools
+
+import (
+	"strings"
+	"testing"
+)
+
+func TestDetectShellCommandIssueFlagsMsysBinaryPaths(t *testing.T) {
+	for _, command := range []string{
+		`for /F %i in ('whoami') do echo %i | "C:\Program Files\Git\usr\bin\head.exe" -1`,
+		`C:\Git\usr\bin\grep.exe pattern file.txt`,
+	} {
+		issue := detectShellCommandIssue(command, "windows")
+		if issue == nil {
+			t.Fatalf("expected MSYS path issue for %q", command)
+		}
+		if issue.Kind != "windows_msys_sandbox" {
+			t.Fatalf("expected windows_msys_sandbox kind, got %q", issue.Kind)
+		}
+		if !strings.Contains(issue.Suggestion, "require_escalated") {
+			t.Fatalf("expected escalation guidance, got %#v", issue)
+		}
+	}
+}
+
+func TestDetectShellCommandIssueFlagsStandaloneCat(t *testing.T) {
+	issue := detectShellCommandIssue(`cat README.md`, "windows")
+	if issue == nil || issue.Kind != "windows_msys_sandbox" {
+		t.Fatalf("expected MSYS sandbox issue for cat, got %#v", issue)
+	}
+}
+
+func TestDetectShellOutputIssueFlagsMsysCreateFileMappingError(t *testing.T) {
+	output := `0 [main] head (3568) C:\Program Files\Git\usr\bin\head.exe: *** fatal error - CreateFileMapping S-1-5-21-3149109338-1484423945-518236903-1001.1, Win32 error 5.  Terminating.`
+	issue := detectShellOutputIssue(output, "windows")
+	if issue == nil || issue.Kind != "windows_msys_sandbox" {
+		t.Fatalf("expected MSYS output issue, got %#v", issue)
+	}
+	if !strings.Contains(issue.Suggestion, "require_escalated") {
+		t.Fatalf("expected escalation guidance, got %#v", issue)
+	}
+}
+
+func TestDetectShellOutputIssueFlagsMsysSignalPipeError(t *testing.T) {
+	output := `0 [main] head (39684) cygheap_user::init: NtSetInformationToken (TokenDefaultDacl), 0xC0000022
+648 [main] head (39684) C:\Program Files\Git\usr\bin\head.exe: *** fatal error - couldn't create signal pipe, Win32 error 5`
+	issue := detectShellOutputIssue(output, "windows")
+	if issue == nil || issue.Kind != "windows_msys_sandbox" {
+		t.Fatalf("expected MSYS output issue, got %#v", issue)
+	}
+}
+
+func TestDetectShellOutputIssueFlagsMsysTerminatingWithMsysMarker(t *testing.T) {
+	output := `1 [main] tail (4321) tail: *** MapViewOfFileEx failed, Win32 error 5.  Terminating.`
+	issue := detectShellOutputIssue(output, "windows")
+	if issue == nil || issue.Kind != "windows_msys_sandbox" {
+		t.Fatalf("expected MSYS output issue, got %#v", issue)
+	}
+}
+
+func TestDetectShellOutputIssueIgnoresNonMsysWin32Error5(t *testing.T) {
+	output := `myapp.exe: unable to open service handle, Win32 error 5 (access denied). Terminating worker.`
+	issue := detectShellOutputIssue(output, "windows")
+	if issue != nil {
+		t.Fatalf("expected no issue for non-MSYS access-denied output, got %#v", issue)
+	}
+}
+
+func TestShellIssueBlockResultMsysCommand(t *testing.T) {
+	result := shellIssueBlockResult(*detectShellCommandIssue(`cat README.md`, "windows"))
+	if result.Status != StatusError {
+		t.Fatalf("status = %q, want error", result.Status)
+	}
+	for _, want := range []string{"[zero] shell issue:", "MSYS/Cygwin", "grep", "read_file", "require_escalated"} {
+		if !strings.Contains(result.Output, want) {
+			t.Fatalf("expected %q in blocked output, got %q", want, result.Output)
+		}
+	}
+	if result.Meta["shell_issue"] != "windows_msys_sandbox" {
+		t.Fatalf("meta shell_issue = %q", result.Meta["shell_issue"])
+	}
+}
+
+func TestMsysProneCommandName(t *testing.T) {
+	if !MsysProneCommandName("HEAD") || MsysProneCommandName("echo") {
+		t.Fatalf("unexpected MsysProneCommandName results")
+	}
+}
+
+// TestDetectShellCommandIssueFlagsExprAndLsConsistently guards against the
+// preflight regex list drifting from the canonical windowsMsysProneNames set
+// (both listed expr and ls as MSYS-prone, but the old regex alternations
+// omitted expr entirely and let ls hit the older windows_shell_syntax branch
+// first, so it never got MSYS-kind guidance).
+func TestDetectShellCommandIssueFlagsExprAndLsConsistently(t *testing.T) {
+	for _, command := range []string{
+		`expr 1 + 1`,
+		`expr.exe 1 + 1`,
+		`ls -la`,
+		`ls`,
+	} {
+		issue := detectShellCommandIssue(command, "windows")
+		if issue == nil || issue.Kind != windowsMsysSandboxKind {
+			t.Fatalf("expected windows_msys_sandbox for %q, got %#v", command, issue)
+		}
+	}
+}
+
+// TestDetectShellCommandIssueIgnoresQuotedMsysMentions guards against
+// treating an MSYS-prone name that only appears inside a quoted argument
+// (e.g. a commit message, a PR comment body, or a doc string discussing the
+// command) as an actual invocation. The preflight check must anchor on the
+// first word of each command segment, not scan the raw text anywhere.
+func TestDetectShellCommandIssueIgnoresQuotedMsysMentions(t *testing.T) {
+	for _, command := range []string{
+		`git commit -m "fix head.exe crash"`,
+		`gh pr comment --body "grep.exe fails under MSYS"`,
+		`echo "log | head is broken on windows"`,
+		`git commit -m "note: | head does not work here"`,
+	} {
+		if issue := detectShellCommandIssue(command, "windows"); issue != nil {
+			t.Fatalf("expected quoted MSYS mention to pass for %q, got %#v", command, issue)
+		}
+	}
+
+	// A real invocation alongside quoted text must still be caught.
+	if issue := detectShellCommandIssue(`echo "not a real command" && head file.txt`, "windows"); issue == nil || issue.Kind != windowsMsysSandboxKind {
+		t.Fatalf("expected real head invocation to still be flagged, got %#v", issue)
+	}
+}
+
+// TestDetectShellCommandIssueIgnoresQuotedMsysPathMentions guards the
+// explicit MSYS-binary-path check the same way as the coreutil-name check:
+// a full usr\bin\ path that only appears inside a quoted argument (e.g. a
+// commit message describing the failure) must not be treated as an
+// invocation, since the path check is now anchored to the first word of each
+// command segment rather than scanned across the raw command text.
+func TestDetectShellCommandIssueIgnoresQuotedMsysPathMentions(t *testing.T) {
+	for _, command := range []string{
+		`git commit -m "C:\Program Files\Git\usr\bin\head.exe fails"`,
+		`gh pr comment --body "C:\Git\usr\bin\grep.exe is blocked"`,
+	} {
+		if issue := detectShellCommandIssue(command, "windows"); issue != nil {
+			t.Fatalf("expected quoted MSYS path mention to pass for %q, got %#v", command, issue)
+		}
+	}
+
+	// A real invocation by full path must still be caught.
+	if issue := detectShellCommandIssue(`C:\Git\usr\bin\grep.exe pattern file.txt`, "windows"); issue == nil || issue.Kind != windowsMsysSandboxKind {
+		t.Fatalf("expected real MSYS path invocation to still be flagged, got %#v", issue)
+	}
+}
+
+// TestDetectShellCommandIssueRespectsCaretEscapedOperators guards against
+// misreading cmd.exe's ^ escape character: `echo ^| head` prints the pipe and
+// "head" as literal text (the caret escapes the pipe so it never splits into
+// a separate `head` invocation), and `echo foo; head` is a single `echo`
+// command with literal arguments since cmd.exe (unlike bash) does not treat
+// ; as a statement separator.
+func TestDetectShellCommandIssueRespectsCaretEscapedOperators(t *testing.T) {
+	for _, command := range []string{
+		`echo ^| head`,
+		`echo ^& head`,
+		`echo foo; head`,
+	} {
+		if issue := detectShellCommandIssue(command, "windows"); issue != nil {
+			t.Fatalf("expected no issue for %q, got %#v", command, issue)
+		}
+	}
+
+	// An unescaped pipe must still split into a real head invocation.
+	if issue := detectShellCommandIssue(`echo foo | head`, "windows"); issue == nil || issue.Kind != windowsMsysSandboxKind {
+		t.Fatalf("expected real head invocation to still be flagged, got %#v", issue)
+	}
+}
+
+// TestDetectShellCommandIssueFlagsRedirectionAttachedToCommand guards against
+// firstCommandWord treating cmd.exe redirection operators as part of the
+// command name. cmd.exe accepts redirection with no separating space
+// (head>out.txt, cat<in.txt), so splitting only on whitespace would return
+// "head>out.txt" as one word and miss the invoked command entirely.
+func TestDetectShellCommandIssueFlagsRedirectionAttachedToCommand(t *testing.T) {
+	for _, command := range []string{
+		`some-command | head>out.txt`,
+		`cat<in.txt`,
+		`grep>matches.txt pattern`,
+	} {
+		issue := detectShellCommandIssue(command, "windows")
+		if issue == nil || issue.Kind != windowsMsysSandboxKind {
+			t.Fatalf("expected windows_msys_sandbox for %q, got %#v", command, issue)
+		}
+	}
+}
+
+// TestDetectShellOutputIssueSignatureOmitsCommandText documents, at the type
+// level, that detectShellOutputIssue can no longer take the command line as
+// evidence: it only accepts the real output. Harmless output must not be
+// flagged, and output that genuinely carries the MSYS failure markers must
+// still be flagged.
+func TestDetectShellOutputIssueSignatureOmitsCommandText(t *testing.T) {
+	if issue := detectShellOutputIssue("hello from bash", "windows"); issue != nil {
+		t.Fatalf("expected no issue for harmless output, got %#v", issue)
+	}
+	output := `0 [main] head (3568) C:\Program Files\Git\usr\bin\head.exe: *** fatal error - CreateFileMapping ..., Win32 error 5.  Terminating.`
+	if issue := detectShellOutputIssue(output, "windows"); issue == nil || issue.Kind != windowsMsysSandboxKind {
+		t.Fatalf("expected real MSYS output to still be flagged, got %#v", issue)
+	}
+}

internal/tools/spill.go

@@ -0,0 +1,145 @@
+package tools
+
+import (
+	"fmt"
+	"os"
+	"path/filepath"
+	"strings"
+	"time"
+
+	"github.com/Gitlawb/zero/internal/redaction"
+	"github.com/Gitlawb/zero/internal/secrets"
+)
+
+// Spill-to-disk for truncated tool output. When a command produces more than
+// its context budget, the tail was previously simply gone — the model's only
+// recourse was re-running the (possibly expensive or non-idempotent) command
+// with a bigger budget. Instead, the full output is written to a cache file
+// whose path is included in the truncation notice, so the model can grep or
+// read_file the remainder. Spilling is best-effort: on any error truncation
+// behaves exactly as before, just without the file hint.
+
+// spillRetention is how long spilled outputs are kept; files older than this
+// are opportunistically removed whenever a new spill happens.
+const spillRetention = 7 * 24 * time.Hour
+
+// spillRootPath returns the spill directory path without creating or checking
+// it. Used by the read-path resolver to recognize spill files as readable.
+func spillRootPath() string {
+	name := "zero-tool-output"
+	if uid := os.Getuid(); uid >= 0 {
+		name = fmt.Sprintf("zero-tool-output-%d", uid)
+	}
+	return filepath.Join(os.TempDir(), name)
+}
+
+// resolveSpillReadPath reports whether requestedPath is a file inside the
+// spill directory and, if so, returns its verified absolute path. Spill files
+// are zero-created, per-uid owned, and already secret-redacted, so letting the
+// scoped read tools open them is what makes the truncation notice's
+// "read_file it" recovery actually work. Symlinks are resolved and the result
+// must still be inside the spill dir, so a planted link cannot smuggle an
+// out-of-scope file through this gate.
+func resolveSpillReadPath(requestedPath string) (string, bool) {
+	if requestedPath == "" || !filepath.IsAbs(requestedPath) {
+		return "", false
+	}
+	root := spillRootPath()
+	cleaned := filepath.Clean(requestedPath)
+	if cleaned == root || !strings.HasPrefix(cleaned, root+string(filepath.Separator)) {
+		return "", false
+	}
+	resolved, err := filepath.EvalSymlinks(cleaned)
+	if err != nil {
+		return "", false
+	}
+	resolvedRoot, err := filepath.EvalSymlinks(root)
+	if err != nil {
+		return "", false
+	}
+	if !strings.HasPrefix(resolved, resolvedRoot+string(filepath.Separator)) {
+		return "", false
+	}
+	return resolved, true
+}
+
+// spillDir returns the per-user spill directory, creating it on first use.
+// Hardened for shared temp dirs (Linux /tmp): the name carries the uid so
+// users cannot collide, and a pre-existing path is only accepted when it is a
+// real directory (not a symlink that would redirect spills) owned by the
+// current user. Any doubt fails the spill — it is best-effort anyway.
+func spillDir() (string, error) {
+	dir := spillRootPath()
+	if err := os.MkdirAll(dir, 0o700); err != nil {
+		return "", err
+	}
+	// MkdirAll follows symlinks and leaves an existing directory untouched, so
+	// verify what is actually at the path.
+	info, err := os.Lstat(dir)
+	if err != nil {
+		return "", err
+	}
+	if !info.IsDir() {
+		return "", fmt.Errorf("spill path %s is not a directory", dir)
+	}
+	if err := checkSpillDirOwner(info); err != nil {
+		return "", err
+	}
+	return dir, nil
+}
+
+// spillTruncatedOutput writes the full pre-truncation output to the spill
+// directory and returns the file path, or "" when spilling fails. Output is
+// scrubbed with the same configured-key redaction the registry applies at the
+// tool boundary PLUS the pattern-based secret scanner (AWS keys, tokens, PEM
+// blocks, JWTs) that bash applies to its model-visible output — a spill runs
+// before that formatter, so without the scan here a spilled file would hold
+// pattern-matched credentials in cleartext that the transcript hides.
+func spillTruncatedOutput(toolName, output string) string {
+	dir, err := spillDir()
+	if err != nil {
+		return ""
+	}
+	sweepSpillDir(dir)
+	prefix := strings.Map(func(r rune) rune {
+		if r >= 'a' && r <= 'z' || r >= 'A' && r <= 'Z' || r >= '0' && r <= '9' || r == '_' || r == '-' {
+			return r
+		}
+		return '-'
+	}, toolName)
+	file, err := os.CreateTemp(dir, prefix+"-*.txt")
+	if err != nil {
+		return ""
+	}
+	defer file.Close()
+	scrubbed := redaction.RedactString(output, redaction.Options{})
+	scrubbed, _ = secrets.Redact(scrubbed)
+	if _, err := file.WriteString(scrubbed); err != nil {
+		_ = os.Remove(file.Name())
+		return ""
+	}
+	return file.Name()
+}
+
+// sweepSpillDir removes spill files older than spillRetention. Best-effort:
+// errors are ignored, the directory is small, and a sweep runs only when a new
+// spill is about to happen anyway.
+func sweepSpillDir(dir string) {
+	entries, err := os.ReadDir(dir)
+	if err != nil {
+		return
+	}
+	cutoff := time.Now().Add(-spillRetention)
+	for _, entry := range entries {
+		if entry.IsDir() {
+			continue
+		}
+		info, err := entry.Info()
+		if err != nil {
+			continue
+		}
+		if info.ModTime().Before(cutoff) {
+			_ = os.Remove(filepath.Join(dir, entry.Name()))
+		}
+	}
+}

internal/tools/spill_hardening_test.go

@@ -0,0 +1,91 @@
+package tools
+
+import (
+	"context"
+	"fmt"
+	"os"
+	"path/filepath"
+	"runtime"
+	"strings"
+	"testing"
+)
+
+func spillDirName() string {
+	if uid := os.Getuid(); uid >= 0 {
+		return fmt.Sprintf("zero-tool-output-%d", uid)
+	}
+	return "zero-tool-output"
+}
+
+func TestSpillDirRejectsSymlink(t *testing.T) {
+	if runtime.GOOS == "windows" {
+		t.Skip("symlink creation needs privileges on Windows")
+	}
+	tmp := t.TempDir()
+	t.Setenv("TMPDIR", tmp)
+	elsewhere := filepath.Join(tmp, "elsewhere")
+	if err := os.Mkdir(elsewhere, 0o700); err != nil {
+		t.Fatal(err)
+	}
+	// An attacker pre-creates the spill path as a symlink to redirect spills.
+	if err := os.Symlink(elsewhere, filepath.Join(tmp, spillDirName())); err != nil {
+		t.Fatal(err)
+	}
+	if path := spillTruncatedOutput("bash", "sensitive output"); path != "" {
+		t.Fatalf("spill must refuse a symlinked directory, wrote %s", path)
+	}
+	entries, _ := os.ReadDir(elsewhere)
+	if len(entries) != 0 {
+		t.Fatalf("nothing may land behind the symlink: %v", entries)
+	}
+}
+
+func TestSpillDirAcceptsOwnedDirectory(t *testing.T) {
+	t.Setenv("TMPDIR", t.TempDir())
+	if path := spillTruncatedOutput("bash", "ok"); path == "" {
+		t.Fatal("spill must work in a clean per-user temp dir")
+	}
+}
+
+func TestResolveSpillReadPathContainment(t *testing.T) {
+	t.Setenv("TMPDIR", t.TempDir())
+	spillPath := spillTruncatedOutput("exec_command", "spilled body")
+	if spillPath == "" {
+		t.Fatal("spill must succeed in a clean temp dir")
+	}
+	resolved, ok := resolveSpillReadPath(spillPath)
+	if !ok || resolved == "" {
+		t.Fatalf("a real spill file must resolve, got ok=%v", ok)
+	}
+	if _, ok := resolveSpillReadPath(filepath.Join(spillRootPath(), "..", "escape.txt")); ok {
+		t.Fatal("path traversal out of the spill dir must be rejected")
+	}
+	if _, ok := resolveSpillReadPath("/etc/hosts"); ok {
+		t.Fatal("paths outside the spill dir must be rejected")
+	}
+	if _, ok := resolveSpillReadPath(spillRootPath()); ok {
+		t.Fatal("the spill dir itself is not a readable file target")
+	}
+	if runtime.GOOS != "windows" {
+		// A symlink planted inside the spill dir pointing outside must not leak.
+		link := filepath.Join(spillRootPath(), "sneaky-link")
+		if err := os.Symlink("/etc/hosts", link); err == nil {
+			if _, ok := resolveSpillReadPath(link); ok {
+				t.Fatal("symlink escaping the spill dir must be rejected")
+			}
+		}
+	}
+}
+
+func TestReadFileCanReadSpillFile(t *testing.T) {
+	t.Setenv("TMPDIR", t.TempDir())
+	spillPath := spillTruncatedOutput("exec_command", "line one\nline two\n")
+	if spillPath == "" {
+		t.Fatal("spill must succeed")
+	}
+	workspace := t.TempDir()
+	result := NewReadFileTool(workspace).Run(context.Background(), map[string]any{"path": spillPath})
+	if result.Status != StatusOK || !strings.Contains(result.Output, "line two") {
+		t.Fatalf("read_file must be able to follow the truncation notice: %s %q", result.Status, result.Output)
+	}
+}

internal/tools/spill_owner_unix.go

@@ -0,0 +1,24 @@
+//go:build !windows
+
+package tools
+
+import (
+	"fmt"
+	"os"
+	"syscall"
+)
+
+// checkSpillDirOwner rejects a spill directory not owned by the current user:
+// on a shared /tmp another user could have pre-created the path and would then
+// control its lifetime (deletion/renaming) even though the 0600 spill files
+// keep their contents private.
+func checkSpillDirOwner(info os.FileInfo) error {
+	stat, ok := info.Sys().(*syscall.Stat_t)
+	if !ok {
+		return nil
+	}
+	if int(stat.Uid) != os.Geteuid() {
+		return fmt.Errorf("spill directory is owned by uid %d, not the current user", stat.Uid)
+	}
+	return nil
+}

internal/tools/spill_owner_windows.go

@@ -0,0 +1,12 @@
+//go:build windows
+
+package tools
+
+import "os"
+
+// checkSpillDirOwner is a no-op on Windows: %TEMP% is per-user by default and
+// there is no portable uid to compare. The Lstat symlink/dir check in spillDir
+// still applies.
+func checkSpillDirOwner(os.FileInfo) error {
+	return nil
+}

internal/tools/spill_test.go

@@ -0,0 +1,109 @@
+package tools
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+	"time"
+)
+
+func TestTruncateExecOutputSpillWritesFullOutput(t *testing.T) {
+	t.Setenv("TMPDIR", t.TempDir())
+	long := strings.Repeat("line of build output\n", 5000) // ~100KB > 40KB budget
+
+	truncated, wasTruncated := truncateExecOutputSpill(long, defaultMaxOutputTokens, "bash")
+	if !wasTruncated {
+		t.Fatal("output over budget must truncate")
+	}
+	if !strings.Contains(truncated, "full output saved to ") {
+		t.Fatalf("truncation notice missing spill path: %q", truncated[:200])
+	}
+	// Extract the path and verify the file holds the complete output.
+	start := strings.Index(truncated, "full output saved to ") + len("full output saved to ")
+	end := strings.Index(truncated[start:], " (grep")
+	spillPath := truncated[start : start+end]
+	content, err := os.ReadFile(spillPath)
+	if err != nil {
+		t.Fatalf("spill file unreadable: %v", err)
+	}
+	if string(content) != long {
+		t.Fatalf("spill file must hold the full output: got %d bytes, want %d", len(content), len(long))
+	}
+}
+
+func TestTruncateExecOutputUnderBudgetUnchanged(t *testing.T) {
+	output, wasTruncated := truncateExecOutputSpill("short output", defaultMaxOutputTokens, "bash")
+	if wasTruncated || output != "short output" {
+		t.Fatalf("under-budget output must pass through untouched: %q", output)
+	}
+}
+
+func TestSweepSpillDirRemovesOnlyOldFiles(t *testing.T) {
+	dir := t.TempDir()
+	oldFile := filepath.Join(dir, "bash-old.txt")
+	newFile := filepath.Join(dir, "bash-new.txt")
+	for _, path := range []string{oldFile, newFile} {
+		if err := os.WriteFile(path, []byte("x"), 0o600); err != nil {
+			t.Fatal(err)
+		}
+	}
+	stale := time.Now().Add(-spillRetention - time.Hour)
+	if err := os.Chtimes(oldFile, stale, stale); err != nil {
+		t.Fatal(err)
+	}
+
+	sweepSpillDir(dir)
+
+	if _, err := os.Stat(oldFile); !os.IsNotExist(err) {
+		t.Fatal("stale spill file must be removed")
+	}
+	if _, err := os.Stat(newFile); err != nil {
+		t.Fatal("fresh spill file must survive the sweep")
+	}
+}
+
+// A spill happens BEFORE bash's model-facing formatter runs its pattern-based
+// secret scan, so the spill itself must scrub pattern-matched credentials —
+// otherwise the file would hold in cleartext exactly what the transcript hides.
+func TestSpillTruncatedOutputScrubsPatternSecrets(t *testing.T) {
+	setTestTempDir(t)
+	githubToken := "ghp_" + strings.Repeat("a", 36)
+	body := "before\nAKIAIOSFODNN7EXAMPLE\n" + githubToken + "\nafter"
+
+	path := spillTruncatedOutput("bash", body)
+	if path == "" {
+		t.Fatal("spill must return a file path")
+	}
+	content, err := os.ReadFile(path)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if strings.Contains(string(content), "AKIAIOSFODNN7EXAMPLE") {
+		t.Fatal("AWS access key reached the spill file in cleartext")
+	}
+	if strings.Contains(string(content), githubToken) {
+		t.Fatal("GitHub token reached the spill file in cleartext")
+	}
+	if !strings.Contains(string(content), "before") || !strings.Contains(string(content), "after") {
+		t.Fatalf("non-secret content must survive the scrub: %q", content)
+	}
+}
+
+func TestSpillTruncatedOutputWritesFile(t *testing.T) {
+	t.Setenv("TMPDIR", t.TempDir())
+	path := spillTruncatedOutput("exec_command", "some output body")
+	if path == "" {
+		t.Fatal("spill must return a file path")
+	}
+	if base := filepath.Base(path); !strings.HasPrefix(base, "exec_command-") {
+		t.Fatalf("spill file name must carry the tool prefix: %s", base)
+	}
+	content, err := os.ReadFile(path)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if string(content) != "some output body" {
+		t.Fatalf("unexpected spill content: %q", content)
+	}
+}

internal/tools/web_fetch.go

@@ -18,12 +18,17 @@ import (
 )
 
 const (
-	defaultWebFetchMaxBytes = 64 * 1024
-	maxWebFetchMaxBytes     = 512 * 1024
-	webFetchTimeout         = 10 * time.Second
+	// Raw-body budgets. HTML responses are converted to markdown before they
+	// reach the model (see web_fetch_markdown.go), so a generous raw budget
+	// does not translate into a generous context cost — conversion typically
+	// shrinks a page by an order of magnitude. 64KiB of raw HTML often held
+	// nothing but a page's <head>, which starved research tasks.
+	defaultWebFetchMaxBytes = 256 * 1024
+	maxWebFetchMaxBytes     = 2 * 1024 * 1024
+	webFetchTimeout         = 30 * time.Second
 	webFetchRedirectLimit   = 5
 	webFetchErrorBodyLimit  = 4 * 1024
-	webFetchPublicOnlyHint  = "web_fetch only supports public remote HTTP/HTTPS URLs. For localhost or private network URLs, use bash with curl so sandbox network permission can apply."
+	webFetchPublicOnlyHint  = "web_fetch only supports public remote HTTP/HTTPS URLs. For localhost or private network URLs, use bash with curl so sandbox network permission can apply"
 )
 
 type webFetchTool struct {
@@ -115,11 +120,17 @@ func newWebFetchToolWithClientAndResolver(client *http.Client, resolver webFetch
 					},
 					"max_bytes": {
 						Type:        "integer",
-						Description: "Maximum response body bytes to return.",
+						Description: "Maximum raw response body bytes to download before conversion.",
 						Default:     defaultWebFetchMaxBytes,
 						Minimum:     intPtr(1),
 						Maximum:     intPtr(maxWebFetchMaxBytes),
 					},
+					"format": {
+						Type:        "string",
+						Description: "auto (default): HTML responses are converted to compact markdown, everything else is returned as-is. raw: never convert. markdown: force conversion.",
+						Enum:        []string{"auto", "raw", "markdown"},
+						Default:     "auto",
+					},
 				},
 				Required:             []string{"url"},
 				AdditionalProperties: false,
@@ -170,6 +181,18 @@ func (tool webFetchTool) run(ctx context.Context, args map[string]any) Result {
 	if err != nil {
 		return errorResult("Error: Invalid arguments for web_fetch: " + err.Error())
 	}
+	format, err := stringArg(args, "format", "auto", false)
+	if err != nil {
+		return errorResult("Error: Invalid arguments for web_fetch: " + err.Error())
+	}
+	format = strings.ToLower(strings.TrimSpace(format))
+	switch format {
+	case "", "auto":
+		format = "auto"
+	case "raw", "markdown":
+	default:
+		return errorResult(`Error: Invalid arguments for web_fetch: format must be "auto", "raw", or "markdown".`)
+	}
 	if err := validateWebFetchURLBeforePermission(rawURL); err != nil {
 		return errorResult("Error: Unsafe URL for web_fetch: " + err.Error())
 	}
@@ -224,14 +247,28 @@ func (tool webFetchTool) run(ctx context.Context, args map[string]any) Result {
 	}
 	body = redactWebFetchText(body)
 
-	output := strings.Join([]string{
+	// HTML responses are converted to compact markdown by default: raw HTML is
+	// mostly markup, so conversion typically shrinks the page by an order of
+	// magnitude and the model reads content instead of boilerplate. format=raw
+	// is the escape hatch for pages the converter mangles.
+	converted := false
+	if format == "markdown" || (format == "auto" && looksLikeHTML(contentType, body)) {
+		if markdown := htmlToMarkdown(body); markdown != "" {
+			body = markdown
+			converted = true
+		}
+	}
+
+	headers := []string{
 		"URL: " + finalURL,
 		"Status: " + webFetchStatusLine(response),
 		"Content-Type: " + firstNonEmptyString(contentType, "unknown"),
 		"Bytes: " + strconv.Itoa(len(body)),
-		"",
-		body,
-	}, "\n")
+	}
+	if converted {
+		headers = append(headers, "Converted: html -> markdown (pass format: \"raw\" for the original HTML)")
+	}
+	output := strings.Join(append(headers, "", body), "\n")
 
 	return Result{
 		Status:    StatusOK,
@@ -243,6 +280,7 @@ func (tool webFetchTool) run(ctx context.Context, args map[string]any) Result {
 			"content_type": contentType,
 			"bytes":        strconv.Itoa(len(body)),
 			"truncated":    strconv.FormatBool(truncated),
+			"converted":    strconv.FormatBool(converted),
 		},
 	}
 }
@@ -310,7 +348,7 @@ func webFetchRedirectPolicy(previous func(*http.Request, []*http.Request) error,
 			return fmt.Errorf("too many redirects: maximum is %d", webFetchRedirectLimit)
 		}
 		if err := validateParsedWebFetchURL(request.Context(), request.URL, resolver); err != nil {
-			return fmt.Errorf("Unsafe redirect URL: %w", err)
+			return fmt.Errorf("unsafe redirect URL: %w", err)
 		}
 		if previous != nil {
 			return previous(request, via)

internal/tools/web_fetch_markdown.go

@@ -0,0 +1,92 @@
+package tools
+
+import (
+	"fmt"
+	"html"
+	"regexp"
+	"strings"
+)
+
+// HTML-to-markdown conversion for web_fetch. Raw HTML is mostly boilerplate —
+// on a typical page well under a tenth of the bytes are readable content — so
+// returning it raw wastes nearly the whole response budget on markup. The
+// converter is deliberately dependency-free (regex passes + stdlib entity
+// unescaping, no HTML parser): the output feeds a language model, not a
+// browser, so best-effort structure (headings, links, list markers) is enough,
+// and the format=raw escape hatch covers pages it mangles.
+
+var (
+	webFetchDropBlockRe = regexp.MustCompile(`(?is)<(script|style|head|noscript|svg|template|iframe)\b[^>]*>.*?</\s*(?:script|style|head|noscript|svg|template|iframe)\s*>`)
+	webFetchCommentRe   = regexp.MustCompile(`(?s)<!--.*?-->`)
+	webFetchAnchorRe    = regexp.MustCompile(`(?is)<a\b[^>]*?href\s*=\s*["']?([^"'\s>]+)["']?[^>]*>(.*?)</\s*a\s*>`)
+	webFetchHeadingRe   = regexp.MustCompile(`(?is)<h([1-6])[^>]*>(.*?)</\s*h[1-6]\s*>`)
+	webFetchListItemRe  = regexp.MustCompile(`(?i)<li\b[^>]*>`)
+	webFetchBreakRe     = regexp.MustCompile(`(?i)<(?:br|hr)\s*/?>`)
+	webFetchBlockTagRe  = regexp.MustCompile(`(?i)</?(?:p|div|section|article|main|header|footer|nav|aside|table|tr|ul|ol|blockquote|pre|figure|form)\b[^>]*>`)
+	webFetchAnyTagRe    = regexp.MustCompile(`(?s)<[^>]*>`)
+	webFetchSpaceRunRe  = regexp.MustCompile(`[ \t]+`)
+	webFetchBlankRunRe  = regexp.MustCompile(`\n{3,}`)
+)
+
+// htmlToMarkdown converts an HTML document to compact markdown-flavored text:
+// headings become #-prefixed lines, anchors become [text](href), list items
+// become "- " bullets, block boundaries become blank lines, all other markup
+// is stripped, and entities are unescaped.
+func htmlToMarkdown(body string) string {
+	text := webFetchCommentRe.ReplaceAllString(body, " ")
+	text = webFetchDropBlockRe.ReplaceAllString(text, " ")
+	text = webFetchHeadingRe.ReplaceAllStringFunc(text, func(match string) string {
+		groups := webFetchHeadingRe.FindStringSubmatch(match)
+		level := int(groups[1][0] - '0')
+		if level < 1 || level > 6 {
+			level = 1
+		}
+		title := strings.TrimSpace(webFetchAnyTagRe.ReplaceAllString(groups[2], " "))
+		return "\n\n" + strings.Repeat("#", level) + " " + title + "\n\n"
+	})
+	text = webFetchAnchorRe.ReplaceAllStringFunc(text, func(match string) string {
+		groups := webFetchAnchorRe.FindStringSubmatch(match)
+		href := groups[1]
+		label := strings.TrimSpace(webFetchAnyTagRe.ReplaceAllString(groups[2], " "))
+		if label == "" {
+			return " "
+		}
+		if strings.HasPrefix(href, "#") || strings.HasPrefix(strings.ToLower(href), "javascript:") {
+			return label
+		}
+		return fmt.Sprintf("[%s](%s)", label, href)
+	})
+	text = webFetchListItemRe.ReplaceAllString(text, "\n- ")
+	text = webFetchBreakRe.ReplaceAllString(text, "\n")
+	text = webFetchBlockTagRe.ReplaceAllString(text, "\n\n")
+	text = webFetchAnyTagRe.ReplaceAllString(text, " ")
+	text = html.UnescapeString(text)
+
+	// Whitespace normalization: collapse space runs, trim line edges, cap blank
+	// runs at one empty line.
+	lines := strings.Split(text, "\n")
+	for i, line := range lines {
+		lines[i] = strings.TrimSpace(webFetchSpaceRunRe.ReplaceAllString(line, " "))
+	}
+	text = strings.Join(lines, "\n")
+	text = webFetchBlankRunRe.ReplaceAllString(text, "\n\n")
+	return strings.TrimSpace(text)
+}
+
+// looksLikeHTML reports whether a response should be treated as an HTML
+// document, by Content-Type first and a body sniff as fallback (servers
+// mislabel HTML as text/plain often enough to matter).
+func looksLikeHTML(contentType, body string) bool {
+	lowered := strings.ToLower(contentType)
+	if strings.Contains(lowered, "text/html") || strings.Contains(lowered, "application/xhtml") {
+		return true
+	}
+	if lowered != "" && !strings.Contains(lowered, "text/plain") {
+		return false
+	}
+	head := strings.ToLower(strings.TrimSpace(body))
+	if len(head) > 512 {
+		head = head[:512]
+	}
+	return strings.Contains(head, "<!doctype html") || strings.Contains(head, "<html")
+}

internal/tools/web_fetch_markdown_test.go

@@ -0,0 +1,127 @@
+package tools
+
+import (
+	"context"
+	"net/http"
+	"strings"
+	"testing"
+)
+
+const sampleHTML = `<!DOCTYPE html>
+<html><head><title>ignored</title><style>body{color:red}</style></head>
+<body>
+<script>var tracking = "noise";</script>
+<h1>Main Title</h1>
+<p>First paragraph with a <a href="https://example.com/docs">docs link</a> and &amp; entity.</p>
+<ul><li>alpha</li><li>beta</li></ul>
+<!-- hidden comment -->
+<h2>Section</h2>
+<p>Body   with   extra   spaces.</p>
+</body></html>`
+
+func TestHTMLToMarkdown(t *testing.T) {
+	markdown := htmlToMarkdown(sampleHTML)
+
+	for _, want := range []string{
+		"# Main Title",
+		"## Section",
+		"[docs link](https://example.com/docs)",
+		"- alpha",
+		"- beta",
+		"with a", // paragraph text survives
+		"& entity",
+	} {
+		if !strings.Contains(markdown, want) {
+			t.Fatalf("markdown missing %q:\n%s", want, markdown)
+		}
+	}
+	for _, banned := range []string{"tracking", "color:red", "hidden comment", "<p>", "</html>", "ignored"} {
+		if strings.Contains(markdown, banned) {
+			t.Fatalf("markdown must not contain %q:\n%s", banned, markdown)
+		}
+	}
+	if strings.Contains(markdown, "extra   spaces") {
+		t.Fatalf("space runs must collapse:\n%s", markdown)
+	}
+	if strings.Contains(markdown, "\n\n\n") {
+		t.Fatalf("blank-line runs must be capped:\n%s", markdown)
+	}
+}
+
+func TestLooksLikeHTML(t *testing.T) {
+	cases := []struct {
+		contentType string
+		body        string
+		want        bool
+	}{
+		{"text/html; charset=utf-8", "anything", true},
+		{"application/xhtml+xml", "anything", true},
+		{"application/json", `{"html":"<html>"}`, false},
+		{"text/plain", "<!doctype html><html>", true}, // mislabeled HTML sniffed
+		{"text/plain", "just text", false},
+		{"", "<html><body>x</body></html>", true},
+	}
+	for _, c := range cases {
+		if got := looksLikeHTML(c.contentType, c.body); got != c.want {
+			t.Errorf("looksLikeHTML(%q, %.20q) = %v, want %v", c.contentType, c.body, got, c.want)
+		}
+	}
+}
+
+func TestWebFetchConvertsHTMLByDefault(t *testing.T) {
+	tool := newWebFetchToolWithClient(webFetchTestClient(func(request *http.Request) (*http.Response, error) {
+		return webFetchTestResponse(request, http.StatusOK, "text/html; charset=utf-8", sampleHTML), nil
+	}))
+
+	result := tool.Run(context.Background(), map[string]any{"url": "https://example.com/page"})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %s: %s", result.Status, result.Output)
+	}
+	if !strings.Contains(result.Output, "# Main Title") || strings.Contains(result.Output, "<h1>") {
+		t.Fatalf("HTML must convert to markdown by default:\n%s", result.Output)
+	}
+	if result.Meta["converted"] != "true" {
+		t.Fatalf("converted meta must be true: %#v", result.Meta)
+	}
+}
+
+func TestWebFetchRawFormatSkipsConversion(t *testing.T) {
+	tool := newWebFetchToolWithClient(webFetchTestClient(func(request *http.Request) (*http.Response, error) {
+		return webFetchTestResponse(request, http.StatusOK, "text/html", sampleHTML), nil
+	}))
+
+	result := tool.Run(context.Background(), map[string]any{"url": "https://example.com/page", "format": "raw"})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %s: %s", result.Status, result.Output)
+	}
+	if !strings.Contains(result.Output, "<h1>Main Title</h1>") {
+		t.Fatalf("format=raw must keep original HTML:\n%s", result.Output)
+	}
+	if result.Meta["converted"] != "false" {
+		t.Fatalf("converted meta must be false: %#v", result.Meta)
+	}
+}
+
+func TestWebFetchLeavesNonHTMLUntouched(t *testing.T) {
+	tool := newWebFetchToolWithClient(webFetchTestClient(func(request *http.Request) (*http.Response, error) {
+		return webFetchTestResponse(request, http.StatusOK, "application/json", `{"key":"<b>value</b>"}`), nil
+	}))
+
+	result := tool.Run(context.Background(), map[string]any{"url": "https://example.com/api"})
+	if result.Status != StatusOK {
+		t.Fatalf("expected ok, got %s: %s", result.Status, result.Output)
+	}
+	if !strings.Contains(result.Output, `{"key":"<b>value</b>"}`) {
+		t.Fatalf("non-HTML must pass through untouched:\n%s", result.Output)
+	}
+}
+
+func TestWebFetchRejectsInvalidFormat(t *testing.T) {
+	tool := newWebFetchToolWithClient(webFetchTestClient(func(request *http.Request) (*http.Response, error) {
+		return webFetchTestResponse(request, http.StatusOK, "text/plain", "x"), nil
+	}))
+	result := tool.Run(context.Background(), map[string]any{"url": "https://example.com", "format": "xml"})
+	if result.Status != StatusError || !strings.Contains(result.Output, "format must be") {
+		t.Fatalf("invalid format must be rejected: %s %s", result.Status, result.Output)
+	}
+}

internal/tools/web_fetch_test.go

@@ -405,7 +405,7 @@ func TestWebFetchToolRejectsUnsafeRedirects(t *testing.T) {
 			if result.Status != StatusError {
 				t.Fatalf("expected redirect safety error, got %s: %s", result.Status, result.Output)
 			}
-			if !strings.Contains(result.Output, "Unsafe redirect URL") {
+			if !strings.Contains(result.Output, "unsafe redirect URL") {
 				t.Fatalf("expected unsafe redirect message, got %q", result.Output)
 			}
 		})

internal/tools/web_search.go

@@ -426,7 +426,6 @@ func stringListArgWebSearch(args map[string]any, key string) (domains []string,
 	if !ok || value == nil {
 		return nil, false, nil
 	}
-	provided = true
 	raw, ok := value.([]any)
 	if !ok {
 		if asString, ok2 := value.([]string); ok2 {

internal/tools/workspace.go

@@ -254,6 +254,14 @@ func scopedReadRoots(workspaceRoot string, scope PathScope) ([]string, error) {
 }
 
 func resolveScopedReadPath(workspaceRoot string, scope PathScope, requestedPath string) (string, string, error) {
+	// Spill files (truncated tool output saved under the per-uid temp dir) are
+	// readable regardless of scope: the truncation notice tells the model to
+	// read_file/grep them, which must actually work. resolveSpillReadPath
+	// verifies containment after symlink resolution, so this cannot be used to
+	// reach anything outside the spill dir.
+	if spillPath, ok := resolveSpillReadPath(requestedPath); ok {
+		return spillPath, spillPath, nil
+	}
 	if requestedPath == "" || !filepath.IsAbs(requestedPath) || scope == nil {
 		return resolveWorkspacePath(workspaceRoot, requestedPath)
 	}

internal/tools/write_file.go

@@ -44,7 +44,7 @@ func (tool writeFileTool) Run(ctx context.Context, args map[string]any) Result {
 	return tool.RunWithOptions(ctx, args, RunOptions{})
 }
 
-func (tool writeFileTool) RunWithOptions(_ context.Context, args map[string]any, options RunOptions) Result {
+func (tool writeFileTool) RunWithOptions(ctx context.Context, args map[string]any, options RunOptions) Result {
 	requestedPath, err := aliasedStringArg(args, []string{"path", "file", "file_path", "filename"}, "", true, false)
 	if err != nil {
 		return errorResult("Error: Invalid arguments for write_file: " + err.Error())
@@ -110,6 +110,10 @@ func (tool writeFileTool) RunWithOptions(_ context.Context, args map[string]any,
 	if err := os.WriteFile(absolutePath, []byte(content), 0o644); err != nil {
 		return errorResult("Error writing file " + relativePath + ": " + err.Error())
 	}
+	// Optional format-on-write (ZERO_FORMAT_ON_WRITE). Must run BEFORE the
+	// FileTracker baseline: recording pre-format content would make the very
+	// next edit look like an external modification and trip the conflict guard.
+	content = maybeFormatWrittenFile(ctx, absolutePath, content)
 	// Baseline the freshly written content so a later edit/overwrite in this
 	// session compares against what is now on disk.
 	newInfo, _ := os.Stat(absolutePath)
@@ -126,6 +130,7 @@ func (tool writeFileTool) RunWithOptions(_ context.Context, args map[string]any,
 		lines++
 	}
 	summary := fmt.Sprintf("%s %s (%d lines).", verb, relativePath, lines)
+	summary += inlineDiagnostics(ctx, options, absolutePath, relativePath)
 	result := okResult(summary)
 	result.ChangedFiles = []string{relativePath}
 	// Card-only preview: a real unified diff (all-green for a create, red/green for

internal/tui/assistant_markdown.go

@@ -8,6 +8,7 @@ import (
 	"unicode/utf8"
 
 	"charm.land/lipgloss/v2"
+	"github.com/charmbracelet/x/ansi"
 )
 
 const (
@@ -237,8 +238,10 @@ func renderAssistantMarkdownPlainText(text string, proseMeasure int, tableMeasur
 }
 
 func stripMarkdownRenderControls(text string) string {
-	text = strings.ReplaceAll(text, markdownBoldStart, "")
-	return strings.ReplaceAll(text, markdownBoldEnd, "")
+	// renderAssistantMarkdownText embeds markdownBoldStart/markdownBoldEnd
+	// markers for prose text, and highlightCodeAuto embeds ANSI color sequences
+	// for fenced code blocks. Strip everything so clipboard text is pure plain text.
+	return ansi.Strip(text)
 }
 
 func styleAssistantMarkdownLine(line string, base lipgloss.Style) string {

internal/tui/autocomplete.go

@@ -188,7 +188,54 @@ func (m model) matchCommandSuggestions(token string) []commandSuggestion {
 	out := matchCommandSuggestionsWithFilter(token, func(command commandDefinition) bool {
 		return command.kind != commandSandboxSetup || m.sandboxSetupCommand != nil
 	})
-	return append(out, m.matchUserCommandSuggestions(token)...)
+	out = append(out, m.matchUserCommandSuggestions(token)...)
+	out = append(out, m.matchSkillSuggestions(token)...)
+	// Each source caps itself, but the merged list must honor the shared cap too
+	// (three sources could otherwise stack up to ~3x the palette bound).
+	if len(out) > maxCommandSuggestions {
+		out = out[:maxCommandSuggestions]
+	}
+	return out
+}
+
+// matchSkillSuggestions returns installed skills whose slash name has the typed
+// prefix, so skills are discoverable and invocable from the palette like user
+// commands. Precedence (builtin > user command > skill) is enforced here by
+// skipping any skill whose name is claimed by a builtin (or alias) or a user
+// command — unlike the builtin/user pair, a shadowed skill row would be dead at
+// dispatch time, so it must not be advertised at all.
+func (m model) matchSkillSuggestions(token string) []commandSuggestion {
+	prefix := strings.ToLower(strings.TrimSpace(strings.TrimPrefix(token, "/")))
+	if prefix == "" || m.loadSkills == nil {
+		return nil
+	}
+	taken := m.takenSlashNames()
+	var out []commandSuggestion
+	// Read through the loader (not the startup snapshot in agentOptions) so the
+	// palette matches what dispatch will actually resolve — a skill installed or
+	// removed mid-session appears/disappears without a restart.
+	for _, skill := range m.installedSkills() {
+		name := skillSlashName(skill.Name)
+		if name == "" || taken[name] || !strings.HasPrefix(name, prefix) {
+			continue
+		}
+		// Two skills whose names collide after lowercasing (e.g. "Deploy" and
+		// "deploy") map to one slash name; only the first — the one dispatch will
+		// run — may be advertised, or the row's description and the executed
+		// instructions silently diverge.
+		taken[name] = true
+		desc := strings.TrimSpace(skill.Description)
+		if desc == "" {
+			desc = "Skill: /" + name
+		} else {
+			desc += " (skill)"
+		}
+		out = append(out, commandSuggestion{Name: "/" + name, Desc: desc})
+		if len(out) >= maxCommandSuggestions {
+			break
+		}
+	}
+	return out
 }
 
 // matchUserCommandSuggestions returns file-sourced /commands whose name has the

internal/tui/command_views.go

@@ -698,3 +698,19 @@ func (m model) debugText() string {
 		}},
 	})
 }
+
+// skillsText is the /skills fallback when NO skills are installed — an install
+// hint. With skills present /skills opens the searchable skill picker instead
+// (see newSkillPicker), matching how /model works.
+func (m model) skillsText() string {
+	return renderCommandOutput(commandOutput{
+		Title:  "Skills",
+		Status: commandStatusInfo,
+		Sections: []commandSection{{
+			Lines: []string{"No skills installed."},
+		}},
+		Hints: []string{
+			"install one: create <skills-dir>/<name>/SKILL.md (see `zero skills`)",
+		},
+	})
+}

internal/tui/commands.go

@@ -46,6 +46,8 @@ const (
 	commandCopy
 	commandExport
 	commandNew
+	commandSkills
+	commandLoop
 	commandUnknown
 )
 
@@ -131,6 +133,13 @@ var commandDefinitions = []commandDefinition{
 		description: "List registered tools.",
 		kind:        commandTools,
 	},
+	{
+		name:        "/skills",
+		usage:       "/skills",
+		group:       commandGroupTools,
+		description: "List installed skills; run one directly with /<skill-name> [args].",
+		kind:        commandSkills,
+	},
 	{
 		name:        "/context",
 		usage:       "/context",
@@ -289,6 +298,13 @@ var commandDefinitions = []commandDefinition{
 		description: "Write the conversation transcript to a file.",
 		kind:        commandExport,
 	},
+	{
+		name:        "/loop",
+		usage:       "/loop [interval] <prompt|/command> | /loop list | /loop stop [id|all]",
+		group:       commandGroupSession,
+		description: "Repeat a prompt or command on an interval (e.g. /loop 5m /babysit-prs), or self-paced when no interval is given.",
+		kind:        commandLoop,
+	},
 	{
 		name:        "/help",
 		usage:       "/help",

internal/tui/composer.go

@@ -280,7 +280,7 @@ func (m model) composerPositionAtMouse(msg tea.MouseMsg) (int, bool) {
 }
 
 func (m model) composerMouseSelectionBlocked() bool {
-	return m.setup.visible || m.providerWizard != nil || m.mcpAddWizard != nil ||
+	return m.transcriptDetailed || m.setup.visible || m.providerWizard != nil || m.mcpAddWizard != nil ||
 		m.mcpManager != nil || m.picker != nil || m.suggestionsActive()
 }
 

internal/tui/file_view_test.go

@@ -143,7 +143,7 @@ func TestFileViewSwapsTranscriptBody(t *testing.T) {
 	m := filesPanelTestModel()
 	m = m.openFileView("internal/tui/sidebar.go")
 
-	items := m.transcriptBodyItems(m.chatColumnWidth(), "")
+	items := m.transcriptBodyItems(m.chatColumnWidth(), "", false)
 	if len(items) != 1 {
 		t.Fatalf("file view should swap the body to a single block item, got %d items", len(items))
 	}
@@ -264,6 +264,55 @@ func TestFileViewFullBodyTruncatesLongFile(t *testing.T) {
 	}
 }
 
+// TestDetailedTranscriptClosesFileView: entering detailed transcript mode
+// closes an active file drill-in so the full transcript body replaces the file
+// content.
+func TestDetailedTranscriptClosesFileView(t *testing.T) {
+	m := filesPanelTestModel()
+	m.altScreen = true
+	m = m.openFileView("web/app.js")
+
+	if !m.fileView.active {
+		t.Fatal("sanity check: openFileView should activate the file view")
+	}
+
+	updated, _ := m.Update(testKeyCtrl('o'))
+	m = updated.(model)
+
+	if m.fileView.active {
+		t.Fatal("detailed transcript should close the file drill-in")
+	}
+	if !m.transcriptDetailed {
+		t.Fatal("Ctrl+O should enter detailed mode")
+	}
+
+	items := m.transcriptBodyItems(m.chatColumnWidth(), "", true)
+	if len(items) <= 1 {
+		t.Fatalf("detailed transcript should show multiple body items, got %d", len(items))
+	}
+}
+
+// TestDetailedTranscriptStaysClosedOnSecondToggle: toggling out and back into
+// detailed mode does not re-open the closed file view.
+func TestDetailedTranscriptStaysClosedOnSecondToggle(t *testing.T) {
+	m := filesPanelTestModel()
+	m.altScreen = true
+	m = m.openFileView("web/app.js")
+
+	updated, _ := m.Update(testKeyCtrl('o'))
+	m = updated.(model)
+
+	updated, _ = m.Update(testKeyCtrl('o'))
+	m = updated.(model)
+
+	if m.fileView.active {
+		t.Fatal("exiting detailed mode must not re-open the file drill-in")
+	}
+	if m.transcriptDetailed {
+		t.Fatal("second Ctrl+O should exit detailed mode")
+	}
+}
+
 // TestFileViewKeysDeferToBlockingModal: with a permission prompt up, Esc and
 // the d/f shortcuts belong to the prompt — the drill-in must not swallow them
 // (Esc exiting the view instead of reaching the prompt's deny handling).

internal/tui/files_panel.go

@@ -372,7 +372,7 @@ func (m model) scrollOffsetForTranscriptRow(rowIndex int) (int, bool) {
 		return 0, false
 	}
 	width := m.chatColumnWidth()
-	items := m.transcriptBodyItems(width, "")
+	items := m.transcriptBodyItems(width, "", false)
 	metrics := measureTranscriptBodyItems(items, m.transcriptBodyHeights)
 	startY := -1
 	for _, span := range metrics.spans {

internal/tui/loop.go

@@ -0,0 +1,693 @@
+package tui
+
+import (
+	"fmt"
+	"regexp"
+	"strconv"
+	"strings"
+	"time"
+
+	tea "charm.land/bubbletea/v2"
+)
+
+// The /loop command repeats a prompt or slash command in one of two modes:
+//
+//   - Fixed interval — `/loop 5m <prompt>` re-fires on a wall-clock cadence.
+//   - Self-paced     — bare `/loop <prompt>` runs a maintenance playbook: the model
+//     does one increment of work each iteration and ends its reply with a control
+//     line (`LOOP: DONE` / `LOOP: CONTINUE [interval]`) that stops the loop or picks
+//     the next wake. With no control line it falls back to an adaptive cadence.
+//
+// Loops are foreground and session-scoped: a tick only fires while the session is
+// idle between turns (it never interrupts a streaming turn), the loop lives with the
+// session that created it, and switching sessions (/new, /resume) stops it. Safety
+// rails cap iterations, age, consecutive failures, and no-progress repetition.
+// Everything here is pure/loop-local; the turn-lifecycle wiring (ticker, launch,
+// continuation) lives in model.go.
+
+type loopMode int
+
+const (
+	loopModeFixed     loopMode = iota // re-fire every fixed interval
+	loopModeSelfPaced                 // model chooses the next wake each iteration
+)
+
+// Safety bounds. A loop is foreground + session-scoped, but these keep a runaway
+// or forgotten loop from burning tokens forever.
+const (
+	loopMinInterval   = 30 * time.Second
+	loopMaxInterval   = 24 * time.Hour
+	loopMaxIterations = 500 // hard ceiling; the loop auto-stops after this many
+
+	// loopMaxAge auto-expires a loop this long after it was created, so a forgotten
+	// loop stops on its own even if it never reaches the iteration cap.
+	loopMaxAge = 24 * time.Hour
+
+	loopSelfPaceMin     = 1 * time.Minute
+	loopSelfPaceMax     = 1 * time.Hour
+	loopSelfPaceDefault = 15 * time.Minute
+
+	// loopDoomThreshold ends a loop after this many consecutive identical final
+	// answers — a self-paced or interval loop that gets stuck repeating the same
+	// no-progress result should stop rather than churn.
+	loopDoomThreshold = 5
+
+	// loopMaxFailures ends a loop after this many consecutive failed iterations, so
+	// a loop whose iterations keep erroring stops instead of churning to the
+	// iteration cap (a failed turn has no final answer, so the doom guard can't see
+	// the lack of progress).
+	loopMaxFailures = 3
+
+	// loopPollInterval is how often the controller checks whether a loop is due.
+	// A loop only fires while idle, so this is a cheap between-turn poll.
+	loopPollInterval = 1 * time.Second
+)
+
+// loopState is one active loop owned by the current session.
+type loopState struct {
+	id        string
+	mode      loopMode
+	prompt    string // the prompt or "/command" re-run each iteration
+	interval  time.Duration
+	iteration int
+	maxIter   int // 0 => loopMaxIterations
+	createdAt time.Time
+	nextRunAt time.Time
+	// lastResult / repeatRun drive doom-loop detection: consecutive identical
+	// final answers past loopDoomThreshold end the loop.
+	lastResult string
+	repeatRun  int
+	// failRun counts consecutive failed iterations; past loopMaxFailures the loop
+	// stops (a failed turn has no final answer for the doom guard to compare).
+	failRun int
+	paused  bool
+}
+
+func (l *loopState) iterationCap() int {
+	if l.maxIter > 0 {
+		return l.maxIter
+	}
+	return loopMaxIterations
+}
+
+// due reports whether the loop should fire at now (active, not paused, past its
+// next-run mark).
+func (l *loopState) due(now time.Time) bool {
+	return !l.paused && !l.nextRunAt.IsZero() && !now.Before(l.nextRunAt)
+}
+
+// cadenceText renders the loop's cadence for status lines.
+func (l *loopState) cadenceText() string {
+	if l.mode == loopModeSelfPaced {
+		return "self-paced"
+	}
+	return "every " + formatLoopDuration(l.interval)
+}
+
+type loopAction int
+
+const (
+	loopActionUsage   loopAction = iota // show usage / list when empty
+	loopActionStart                     // start a new loop
+	loopActionStop                      // stop one by id
+	loopActionStopAll                   // stop every loop
+	loopActionList                      // list active loops
+)
+
+// loopCommand is the parsed form of a `/loop ...` invocation (the text after the
+// command name).
+type loopCommand struct {
+	action   loopAction
+	targetID string
+	mode     loopMode
+	prompt   string
+	interval time.Duration
+	// note carries a non-fatal advisory (e.g. an interval that was clamped) to
+	// surface back to the user.
+	note string
+}
+
+var loopIntervalRe = regexp.MustCompile(`^(\d+)(s|m|h|d)$`)
+
+// loopControlRe matches a self-paced loop's control line at the start of a line,
+// tolerant of a leading markdown bullet/quote marker and any trailing note after
+// the keyword: "LOOP: DONE", "- loop: continue", "LOOP: CONTINUE 10m — still tidying".
+// The leading ^ anchor keeps a mid-sentence mention ("I'll emit LOOP: DONE when…")
+// from matching; parseLoopControl normalizes CRLF first so `$` can reach line end.
+var loopControlRe = regexp.MustCompile(`(?im)^[^\S\r\n]*(?:[-*>•][^\S\r\n]*)?loop:[^\S\r\n]*(done|continue)\b(?:[^\S\r\n]+(\d+[smhd])\b)?.*$`)
+
+// parseLoopCommand interprets the arguments to `/loop`. Forms:
+//
+//	(empty)                 -> usage/list
+//	list | status          -> list active loops
+//	stop | stop all | stop <id>
+//	<interval> <prompt>     -> fixed-interval loop (e.g. "5m /babysit")
+//	<prompt> every <interval>
+//	<prompt>                -> self-paced loop (no interval)
+func parseLoopCommand(args string) loopCommand {
+	trimmed := strings.TrimSpace(args)
+	if trimmed == "" {
+		return loopCommand{action: loopActionUsage}
+	}
+
+	first := strings.ToLower(strings.Fields(trimmed)[0])
+	switch first {
+	case "list", "status", "ls":
+		return loopCommand{action: loopActionList}
+	case "stop", "cancel", "kill":
+		rest := strings.TrimSpace(trimmed[len(strings.Fields(trimmed)[0]):])
+		if rest == "" || strings.EqualFold(rest, "all") {
+			if strings.EqualFold(rest, "all") {
+				return loopCommand{action: loopActionStopAll}
+			}
+			// bare "stop" stops all when ambiguous; the caller may special-case a
+			// single active loop.
+			return loopCommand{action: loopActionStop, targetID: ""}
+		}
+		return loopCommand{action: loopActionStop, targetID: strings.TrimSpace(rest)}
+	}
+
+	// Leading interval token: "5m <prompt>".
+	if fields := strings.Fields(trimmed); len(fields) >= 2 {
+		if d, ok := parseLoopInterval(fields[0]); ok {
+			prompt := strings.TrimSpace(trimmed[len(fields[0]):])
+			cmd := loopCommand{action: loopActionStart, mode: loopModeFixed, prompt: prompt}
+			cmd.interval, cmd.note = clampLoopInterval(d)
+			return cmd
+		}
+	}
+
+	// Trailing "every <interval>".
+	if prompt, d, ok := splitTrailingEvery(trimmed); ok {
+		cmd := loopCommand{action: loopActionStart, mode: loopModeFixed, prompt: prompt}
+		cmd.interval, cmd.note = clampLoopInterval(d)
+		return cmd
+	}
+
+	// No interval anywhere -> self-paced.
+	return loopCommand{action: loopActionStart, mode: loopModeSelfPaced, prompt: trimmed}
+}
+
+// splitTrailingEvery matches "<prompt> every <interval>" (the interval must be a
+// real duration token so "check every PR" is not mis-parsed as a cadence).
+func splitTrailingEvery(input string) (string, time.Duration, bool) {
+	fields := strings.Fields(input)
+	if len(fields) < 3 {
+		return "", 0, false
+	}
+	last := fields[len(fields)-1]
+	prev := fields[len(fields)-2]
+	if !strings.EqualFold(prev, "every") {
+		return "", 0, false
+	}
+	d, ok := parseLoopInterval(last)
+	if !ok {
+		return "", 0, false
+	}
+	prompt := strings.TrimSpace(strings.Join(fields[:len(fields)-2], " "))
+	if prompt == "" {
+		return "", 0, false
+	}
+	return prompt, d, true
+}
+
+// parseLoopInterval parses a bare "<N><unit>" token (s/m/h/d) into a duration.
+func parseLoopInterval(token string) (time.Duration, bool) {
+	m := loopIntervalRe.FindStringSubmatch(strings.ToLower(strings.TrimSpace(token)))
+	if m == nil {
+		return 0, false
+	}
+	n, err := strconv.Atoi(m[1])
+	if err != nil || n <= 0 {
+		return 0, false
+	}
+	unit := map[string]time.Duration{
+		"s": time.Second,
+		"m": time.Minute,
+		"h": time.Hour,
+		"d": 24 * time.Hour,
+	}[m[2]]
+	return time.Duration(n) * unit, true
+}
+
+// clampLoopInterval bounds an interval to [loopMinInterval, loopMaxInterval] and
+// returns an advisory note when it was adjusted.
+func clampLoopInterval(d time.Duration) (time.Duration, string) {
+	switch {
+	case d < loopMinInterval:
+		return loopMinInterval, fmt.Sprintf("interval raised to the %s minimum", formatLoopDuration(loopMinInterval))
+	case d > loopMaxInterval:
+		return loopMaxInterval, fmt.Sprintf("interval lowered to the %s maximum", formatLoopDuration(loopMaxInterval))
+	default:
+		return d, ""
+	}
+}
+
+// loopTickMsg is the between-turn poll that fires a due loop while the session is
+// idle. seq invalidates a tick left over from before the loop set changed.
+type loopTickMsg struct{ seq int }
+
+// handleLoopCommand dispatches a parsed `/loop ...` invocation.
+func (m model) handleLoopCommand(args string) (model, tea.Cmd) {
+	cmd := parseLoopCommand(args)
+	switch cmd.action {
+	case loopActionList:
+		return m.appendLoopSystem(m.loopListText()), nil
+	case loopActionUsage:
+		if len(m.loops) > 0 {
+			return m.appendLoopSystem(m.loopListText()), nil
+		}
+		return m.appendLoopSystem(loopUsageText()), nil
+	case loopActionStop:
+		return m.stopLoop(cmd.targetID)
+	case loopActionStopAll:
+		return m.stopAllLoops()
+	case loopActionStart:
+		return m.startLoop(cmd)
+	}
+	return m, nil
+}
+
+// startLoop registers a new loop and (re)starts the idle poll ticker. The first
+// iteration fires on the next idle tick (nextRunAt = now).
+func (m model) startLoop(cmd loopCommand) (model, tea.Cmd) {
+	if reason, ok := m.validateLoopTarget(cmd.prompt); !ok {
+		return m.appendLoopSystem(reason), nil
+	}
+	// A /command loop must be fixed-interval. Self-paced mode drives a free-form
+	// goal through the playbook + LOOP: control line (see firePrompt), which a
+	// custom command's expanded template never carries — so a bare self-paced
+	// /command would silently degrade to adaptive polling with no way to report
+	// done. Require an explicit interval for command loops instead.
+	if cmd.mode == loopModeSelfPaced && strings.HasPrefix(strings.TrimSpace(cmd.prompt), "/") {
+		return m.appendLoopSystem("A /command loop needs an interval — e.g. /loop 5m " + strings.Fields(cmd.prompt)[0] + ". Self-paced mode (no interval) is for a free-form goal."), nil
+	}
+	m.loopCounter++
+	id := "L" + strconv.Itoa(m.loopCounter)
+	loop := &loopState{
+		id:        id,
+		mode:      cmd.mode,
+		prompt:    cmd.prompt,
+		interval:  cmd.interval,
+		createdAt: m.now(),
+		nextRunAt: m.now(), // fire the first iteration on the next idle tick
+	}
+	m.loops = append(m.loops, loop)
+	note := ""
+	if cmd.note != "" {
+		note = " (" + cmd.note + ")"
+	}
+	m = m.appendLoopSystem(fmt.Sprintf(
+		"Loop %s started — %s%s. Runs while this session is idle; stop with /loop stop %s.",
+		id, loop.cadenceText(), note, id))
+	return m.ensureLoopTick()
+}
+
+// stopLoop stops one loop by id, or the single active loop when the id is omitted.
+func (m model) stopLoop(id string) (model, tea.Cmd) {
+	if len(m.loops) == 0 {
+		return m.appendLoopSystem("No active loops."), nil
+	}
+	if id == "" {
+		if len(m.loops) == 1 {
+			return m.removeLoop(m.loops[0].id, "Loop "+m.loops[0].id+" stopped.").ensureLoopTick()
+		}
+		return m.appendLoopSystem("Multiple loops active — use /loop stop <id> or /loop stop all.\n" + m.loopListText()), nil
+	}
+	if m.findLoop(id) == nil {
+		return m.appendLoopSystem("No loop " + id + ". " + m.loopListText()), nil
+	}
+	// removeLoop stops the poll ticker; re-arm it so any OTHER loops that remain
+	// keep firing (ensureLoopTick no-ops when this was the last loop).
+	return m.removeLoop(id, "Loop "+id+" stopped.").ensureLoopTick()
+}
+
+func (m model) stopAllLoops() (model, tea.Cmd) {
+	if len(m.loops) == 0 {
+		return m.appendLoopSystem("No active loops."), nil
+	}
+	n := len(m.loops)
+	m.loops = nil
+	m.loopSeq++ // invalidate the pending poll tick
+	m.loopTicking = false
+	return m.appendLoopSystem(fmt.Sprintf("Stopped all %d loop(s).", n)), nil
+}
+
+// fireDueLoopIfIdle fires the earliest due loop when the session is idle. Called
+// from the poll tick; a no-op while a turn, modal, or queued user message is
+// pending (the loop simply waits for the next idle tick).
+func (m model) fireDueLoopIfIdle() (model, tea.Cmd) {
+	if m.loopBusy() || len(m.loops) == 0 {
+		return m, nil
+	}
+	now := m.now()
+	var due *loopState
+	for _, l := range m.loops {
+		if l.due(now) && (due == nil || l.nextRunAt.Before(due.nextRunAt)) {
+			due = l
+		}
+	}
+	if due == nil {
+		return m, nil
+	}
+	m.activeLoopID = due.id
+	due.nextRunAt = time.Time{} // mark running
+	return m.fireLoopPrompt(due.firePrompt())
+}
+
+// firePrompt is the text actually sent for this iteration. A self-paced plain
+// prompt is wrapped in the maintenance playbook so the model knows to do one
+// increment of work and end with a LOOP: control line; a /command or a
+// fixed-interval prompt is sent verbatim.
+func (l *loopState) firePrompt() string {
+	if l.mode == loopModeSelfPaced && !strings.HasPrefix(strings.TrimSpace(l.prompt), "/") {
+		return loopSelfPacePlaybook(l)
+	}
+	return l.prompt
+}
+
+// fireLoopPrompt runs a loop's prompt as a turn — a custom /command if it resolves
+// to one, otherwise a plain prompt.
+func (m model) fireLoopPrompt(prompt string) (model, tea.Cmd) {
+	if strings.HasPrefix(strings.TrimSpace(prompt), "/") {
+		if next, teaCmd, ok := m.handleUserCommand(prompt); ok {
+			return next, teaCmd
+		}
+	}
+	return m.launchPrompt(prompt)
+}
+
+// advanceLoop updates a loop after one of its iterations completes: it folds the
+// result into the failure/no-progress guards, honors a self-paced DONE/next-wake
+// control line, enforces the iteration and age caps, then schedules the next wake
+// (or stops the loop).
+func (m model) advanceLoop(id, finalAnswer string, runErr error) model {
+	l := m.findLoop(id)
+	if l == nil {
+		return m // stopped mid-run
+	}
+	l.iteration++
+	if runErr != nil {
+		l.failRun++
+	} else {
+		l.failRun = 0
+	}
+	res := strings.TrimSpace(finalAnswer)
+	if res != "" && res == l.lastResult {
+		l.repeatRun++
+	} else {
+		l.repeatRun = 0
+	}
+	l.lastResult = res
+
+	// Only self-paced iterations run the playbook, so only they can carry a control
+	// line; a failed turn has no trustworthy final answer to read one from.
+	done, wake := false, time.Duration(0)
+	if l.mode == loopModeSelfPaced && runErr == nil {
+		done, wake = parseLoopControl(finalAnswer)
+	}
+
+	switch {
+	case done:
+		return m.removeLoop(id, fmt.Sprintf("Loop %s finished — the task reported done after %d iteration(s).", id, l.iteration))
+	case l.failRun >= loopMaxFailures:
+		return m.removeLoop(id, fmt.Sprintf("Loop %s stopped after %d consecutive failed iterations.", id, l.failRun))
+	case l.iteration >= l.iterationCap():
+		return m.removeLoop(id, fmt.Sprintf("Loop %s stopped at its %d-iteration cap.", id, l.iteration))
+	case !l.createdAt.IsZero() && m.now().Sub(l.createdAt) >= loopMaxAge:
+		return m.removeLoop(id, fmt.Sprintf("Loop %s expired after %s.", id, formatLoopDuration(loopMaxAge)))
+	case l.repeatRun >= loopDoomThreshold:
+		return m.removeLoop(id, fmt.Sprintf("Loop %s stopped — %d identical results in a row (no progress).", id, l.repeatRun))
+	}
+
+	if l.mode == loopModeSelfPaced {
+		// A model-chosen next wake (from the control line) wins; otherwise fall back to
+		// an adaptive cadence that checks back often while work is fresh and settles to
+		// a slow heartbeat as iterations accrue.
+		delay := wake
+		if delay <= 0 {
+			delay = adaptiveSelfPaceDelay(l.iteration)
+		}
+		l.nextRunAt = m.now().Add(clampSelfPaceDelay(delay))
+	} else {
+		l.nextRunAt = m.now().Add(l.interval)
+	}
+	return m
+}
+
+// parseLoopControl reads a self-paced loop's control signal from the model's final
+// answer. The last control line wins (the model may reference the protocol before
+// emitting it). Returns done=true to stop the loop, or a next-wake hint (0 = none).
+func parseLoopControl(answer string) (done bool, wake time.Duration) {
+	// Normalize CRLF/CR so the line-end anchor can reach `$` (Go's (?m)$ sits right
+	// before `\n`, and the horizontal-whitespace class can't consume a stray `\r`).
+	norm := strings.ReplaceAll(answer, "\r\n", "\n")
+	norm = strings.ReplaceAll(norm, "\r", "\n")
+	matches := loopControlRe.FindAllStringSubmatch(norm, -1)
+	if len(matches) == 0 {
+		return false, 0
+	}
+	last := matches[len(matches)-1]
+	if strings.EqualFold(last[1], "done") {
+		return true, 0
+	}
+	if last[2] != "" {
+		if d, ok := parseLoopInterval(last[2]); ok {
+			return false, d
+		}
+	}
+	return false, 0
+}
+
+// loopSelfPacePlaybook wraps a self-paced loop's goal in the maintenance-loop
+// instructions the model follows each iteration: do one real increment of work,
+// then emit a single control line so the harness knows whether to stop or when to
+// wake next — the self-paced maintenance playbook the feature ships with.
+func loopSelfPacePlaybook(l *loopState) string {
+	var b strings.Builder
+	b.WriteString("You are running in a self-paced maintenance loop (iteration ")
+	b.WriteString(strconv.Itoa(l.iteration + 1))
+	b.WriteString("). Do one concrete increment of work toward the goal below — actual work, not just a plan — then stop and report.\n\n")
+	b.WriteString("Goal: ")
+	b.WriteString(strings.TrimSpace(l.prompt))
+	b.WriteString("\n\nEnd your reply with a control line — on its own line, starting with \"LOOP:\":\n")
+	b.WriteString("  LOOP: DONE          — the goal is fully met; the loop stops.\n")
+	b.WriteString("  LOOP: CONTINUE      — more to do; the loop wakes again automatically.\n")
+	b.WriteString("  LOOP: CONTINUE 10m  — more to do; wake after the delay you name (e.g. 5m, 1h).\n")
+	b.WriteString("A short note after the keyword is fine. Report DONE only when the goal is genuinely complete — the loop trusts that signal to stop.")
+	return b.String()
+}
+
+// adaptiveSelfPaceDelay picks a self-paced loop's next-wake by iteration: frequent
+// early (work is actively progressing), slowing to a heartbeat as it matures.
+func adaptiveSelfPaceDelay(iteration int) time.Duration {
+	switch {
+	case iteration < 2:
+		return 2 * time.Minute
+	case iteration < 5:
+		return 5 * time.Minute
+	case iteration < 10:
+		return 15 * time.Minute
+	default:
+		return 30 * time.Minute
+	}
+}
+
+func (m model) findLoop(id string) *loopState {
+	for _, l := range m.loops {
+		if l.id == id {
+			return l
+		}
+	}
+	return nil
+}
+
+func (m model) removeLoop(id, note string) model {
+	kept := make([]*loopState, 0, len(m.loops))
+	for _, l := range m.loops {
+		if l.id != id {
+			kept = append(kept, l)
+		}
+	}
+	m.loops = kept
+	m.loopSeq++ // invalidate the pending poll tick; a fresh one is scheduled if loops remain
+	m.loopTicking = false
+	if note != "" {
+		m = m.appendLoopSystem(note)
+	}
+	return m
+}
+
+// ensureLoopTick starts the idle poll ticker if it is not already running.
+func (m model) ensureLoopTick() (model, tea.Cmd) {
+	if m.loopTicking || len(m.loops) == 0 {
+		return m, nil
+	}
+	m.loopTicking = true
+	return m, m.scheduleLoopTick()
+}
+
+func (m model) scheduleLoopTick() tea.Cmd {
+	seq := m.loopSeq
+	return tea.Tick(loopPollInterval, func(time.Time) tea.Msg {
+		return loopTickMsg{seq: seq}
+	})
+}
+
+// loopBusy reports whether the session is too busy for a loop to fire — a turn,
+// queued user message, compaction, or ANY blocking modal (permission/askUser/spec
+// review, provider or MCP wizard, MCP manager, or an open picker) is in flight.
+// The modal check is load-bearing: a loop must not launch a run behind an open
+// /resume picker, or that run would complete into whatever session the user then
+// switches to. Mirrors launchQueuedMessageIfReady so loops and queued prompts never
+// collide.
+func (m model) loopBusy() bool {
+	return m.pending || m.exiting || m.compactInFlight || m.hasQueuedMessage() || !m.noBlockingModal()
+}
+
+func (m model) appendLoopSystem(text string) model {
+	m.transcript = reduceTranscript(m.transcript, transcriptAction{kind: actionAppendSystem, text: text})
+	return m
+}
+
+// loopActive reports whether any loop is running (for the footer + leave warning).
+func (m model) loopActive() bool { return len(m.loops) > 0 }
+
+// loopFooterSummary renders the persistent "N loops · next 3:05pm" footer segment,
+// or "" when no loops are active.
+func (m model) loopFooterSummary() string {
+	if len(m.loops) == 0 {
+		return ""
+	}
+	var next time.Time
+	for _, l := range m.loops {
+		if l.nextRunAt.IsZero() || l.paused {
+			continue
+		}
+		if next.IsZero() || l.nextRunAt.Before(next) {
+			next = l.nextRunAt
+		}
+	}
+	label := fmt.Sprintf("%d loop", len(m.loops))
+	if len(m.loops) != 1 {
+		label += "s"
+	}
+	if next.IsZero() {
+		return label
+	}
+	return label + " · next " + next.Format("3:04pm")
+}
+
+func (m model) loopListText() string {
+	if len(m.loops) == 0 {
+		return "No active loops."
+	}
+	var b strings.Builder
+	b.WriteString("Active loops:\n")
+	now := m.now()
+	for _, l := range m.loops {
+		when := "running"
+		if !l.nextRunAt.IsZero() {
+			if d := l.nextRunAt.Sub(now); d > 0 {
+				when = "next in " + formatLoopDuration(d.Round(time.Second))
+			} else {
+				when = "due"
+			}
+		}
+		b.WriteString(fmt.Sprintf("  %s · %s · iter %d · %s · %s\n",
+			l.id, l.cadenceText(), l.iteration, when, truncateLoopPrompt(l.prompt)))
+	}
+	b.WriteString("Stop with /loop stop <id> or /loop stop all.")
+	return b.String()
+}
+
+func loopUsageText() string {
+	return "Repeat a prompt or command:\n" +
+		"  /loop 5m /babysit-prs   — run a command every 5 minutes\n" +
+		"  /loop watch CI every 2m — trailing interval form\n" +
+		"  /loop keep tidying docs — self-paced (the model picks its own cadence and stops when done)\n" +
+		"  /loop list              — show active loops\n" +
+		"  /loop stop [id|all]     — stop a loop\n" +
+		"Loops run while this session is idle and stop when you close it."
+}
+
+// validateLoopTarget rejects looping a built-in command or an unresolved custom
+// command — only a prompt or an existing custom /command (.zero/commands/<name>.md)
+// may loop. Resolving custom commands up front stops a typo'd `/loop 5m /nope` from
+// being scheduled and then re-sent as literal text on every tick.
+func (m model) validateLoopTarget(prompt string) (string, bool) {
+	p := strings.TrimSpace(prompt)
+	if p == "" {
+		return "Nothing to loop — give a prompt or a /command.", false
+	}
+	if strings.HasPrefix(p, "/") {
+		switch parseCommand(p).kind {
+		case commandPrompt:
+			// A bare "/" or similar that isn't a command — treat as a plain prompt.
+		case commandUnknown:
+			name, _ := splitUserCommand(p)
+			if _, ok := m.lookupUserCommand(name); !ok {
+				return "No command /" + name + " — a loop can only run a prompt or an existing custom /command.", false
+			}
+		default:
+			return "A loop can only run a prompt or a custom /command, not a built-in like " + strings.Fields(p)[0] + ".", false
+		}
+	}
+	return "", true
+}
+
+// clearLoopsForSessionSwitch drops every active loop and its ticker. Loops belong to
+// the session that created them; carrying them across /new or /resume would fire the
+// old session's prompt into an unrelated conversation. Returns the count cleared so
+// the caller can note it. Pure state reset — no transcript writes.
+func (m model) clearLoopsForSessionSwitch() (model, int) {
+	n := len(m.loops)
+	if n == 0 {
+		return m, 0
+	}
+	m.loops = nil
+	m.activeLoopID = ""
+	m.loopTicking = false
+	m.loopSeq++ // invalidate any pending poll tick
+	return m, n
+}
+
+func truncateLoopPrompt(prompt string) string {
+	const max = 48
+	p := strings.TrimSpace(strings.ReplaceAll(prompt, "\n", " "))
+	runes := []rune(p)
+	if len(runes) <= max {
+		return p
+	}
+	return strings.TrimSpace(string(runes[:max])) + "…"
+}
+
+// clampSelfPaceDelay bounds a model-chosen self-paced delay to the self-pace
+// band, defaulting an unset/invalid value.
+func clampSelfPaceDelay(d time.Duration) time.Duration {
+	switch {
+	case d <= 0:
+		return loopSelfPaceDefault
+	case d < loopSelfPaceMin:
+		return loopSelfPaceMin
+	case d > loopSelfPaceMax:
+		return loopSelfPaceMax
+	default:
+		return d
+	}
+}
+
+// formatLoopDuration renders a duration compactly (90s, 5m, 2h, 1d) for status
+// text, preferring the largest whole unit.
+func formatLoopDuration(d time.Duration) string {
+	switch {
+	case d%(24*time.Hour) == 0 && d >= 24*time.Hour:
+		return strconv.Itoa(int(d/(24*time.Hour))) + "d"
+	case d%time.Hour == 0 && d >= time.Hour:
+		return strconv.Itoa(int(d/time.Hour)) + "h"
+	case d%time.Minute == 0 && d >= time.Minute:
+		return strconv.Itoa(int(d/time.Minute)) + "m"
+	default:
+		return strconv.Itoa(int(d/time.Second)) + "s"
+	}
+}

internal/tui/loop_controller_test.go

@@ -0,0 +1,460 @@
+package tui
+
+import (
+	"context"
+	"errors"
+	"strings"
+	"testing"
+	"time"
+
+	tea "charm.land/bubbletea/v2"
+
+	"github.com/Gitlawb/zero/internal/usercommands"
+)
+
+func loopTestModel(t *testing.T, now time.Time) model {
+	t.Helper()
+	m := newModel(context.Background(), Options{})
+	m.now = func() time.Time { return now }
+	return m
+}
+
+func startFixedLoop(m model, prompt string, interval time.Duration) model {
+	m, _ = m.startLoop(loopCommand{action: loopActionStart, mode: loopModeFixed, prompt: prompt, interval: interval})
+	return m
+}
+
+func TestStartLoopRegistersAndSchedules(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "check the build", 5*time.Minute)
+
+	if len(m.loops) != 1 {
+		t.Fatalf("want 1 loop, got %d", len(m.loops))
+	}
+	l := m.loops[0]
+	if l.mode != loopModeFixed || l.interval != 5*time.Minute || l.prompt != "check the build" {
+		t.Fatalf("unexpected loop %+v", l)
+	}
+	// First iteration is due immediately (nextRunAt == now).
+	if !l.due(now) {
+		t.Errorf("first iteration should be due at creation")
+	}
+	if !m.loopTicking {
+		t.Errorf("starting a loop should start the poll ticker")
+	}
+}
+
+func TestAdvanceLoopSchedulesNextWake(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "poll", 5*time.Minute)
+	id := m.loops[0].id
+
+	m = m.advanceLoop(id, "did some work", nil)
+	l := m.loops[0]
+	if l.iteration != 1 {
+		t.Fatalf("iteration = %d, want 1", l.iteration)
+	}
+	if !l.nextRunAt.Equal(now.Add(5 * time.Minute)) {
+		t.Fatalf("nextRunAt = %v, want +5m", l.nextRunAt)
+	}
+}
+
+func TestSelfPacedUsesModelDelay(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m, _ = m.startLoop(loopCommand{action: loopActionStart, mode: loopModeSelfPaced, prompt: "keep tidying"})
+	id := m.loops[0].id
+	// The model carries its chosen next wake in the reply's control line.
+	m = m.advanceLoop(id, "made progress\nLOOP: CONTINUE 10m", nil)
+	if len(m.loops) == 0 {
+		t.Fatal("a CONTINUE control line should keep the loop running")
+	}
+	if !m.loops[0].nextRunAt.Equal(now.Add(10 * time.Minute)) {
+		t.Fatalf("self-paced nextRunAt = %v, want +10m", m.loops[0].nextRunAt)
+	}
+}
+
+func TestLoopDoneStops(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m, _ = m.startLoop(loopCommand{action: loopActionStart, mode: loopModeSelfPaced, prompt: "finish the task"})
+	id := m.loops[0].id
+	m = m.advanceLoop(id, "all done\nLOOP: DONE", nil)
+	if len(m.loops) != 0 {
+		t.Fatalf("loop should stop when the task reports done, %d remain", len(m.loops))
+	}
+}
+
+func TestFixedLoopIgnoresControlLine(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "watch CI", time.Minute)
+	id := m.loops[0].id
+	// A fixed-interval loop never runs the playbook, so a stray control line in the
+	// output is not a completion signal — it keeps its wall-clock cadence.
+	m = m.advanceLoop(id, "LOOP: DONE", nil)
+	if len(m.loops) != 1 {
+		t.Fatal("a fixed-interval loop should ignore a control line and keep running")
+	}
+	if !m.loops[0].nextRunAt.Equal(now.Add(time.Minute)) {
+		t.Fatalf("fixed nextRunAt = %v, want +1m", m.loops[0].nextRunAt)
+	}
+}
+
+func TestLoopExpiresByAge(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "forgotten", time.Minute)
+	id := m.loops[0].id
+	m.loops[0].createdAt = now.Add(-loopMaxAge - time.Hour) // created past the age cap
+	m = m.advanceLoop(id, "still going", nil)
+	if len(m.loops) != 0 {
+		t.Fatal("a loop past its age cap should auto-expire")
+	}
+}
+
+func TestLoopDoomGuardStops(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "spin", time.Minute)
+	id := m.loops[0].id
+	for i := 0; i < loopDoomThreshold+1 && len(m.loops) > 0; i++ {
+		m = m.advanceLoop(id, "identical answer", nil)
+	}
+	if len(m.loops) != 0 {
+		t.Fatalf("doom guard should stop a loop repeating identical results")
+	}
+}
+
+func TestLoopIterationCapStops(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "bounded", time.Minute)
+	m.loops[0].maxIter = 3
+	id := m.loops[0].id
+	for i := 0; i < 3 && len(m.loops) > 0; i++ {
+		m = m.advanceLoop(id, "answer "+string(rune('a'+i)), nil) // distinct answers so doom doesn't fire
+	}
+	if len(m.loops) != 0 {
+		t.Fatalf("loop should stop at its iteration cap")
+	}
+}
+
+func TestStopLoopByID(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "a", time.Minute)
+	m = startFixedLoop(m, "b", time.Minute)
+	id := m.loops[0].id
+	m, cmd := m.stopLoop(id)
+	if len(m.loops) != 1 || m.findLoop(id) != nil {
+		t.Fatalf("stopLoop(%s) should remove exactly that loop", id)
+	}
+	// Stopping one of several loops must keep the poll ticker alive for the rest:
+	// removeLoop kills the ticker, so stopLoop has to re-arm it (jatmn).
+	if cmd == nil || !m.loopTicking {
+		t.Fatal("stopping one of several loops must leave the remaining loop scheduled (ticker re-armed)")
+	}
+	// Bare stop with a single active loop stops that sole loop (parseLoopCommand
+	// leaves the target empty and stopLoop special-cases the one-loop case).
+	soleID := m.loops[0].id
+	m, _ = m.stopLoop("")
+	if len(m.loops) != 0 || m.findLoop(soleID) != nil {
+		t.Fatalf("bare stop should remove the sole active loop")
+	}
+	m = startFixedLoop(m, "c", time.Minute)
+	m = startFixedLoop(m, "d", time.Minute)
+	m, _ = m.stopAllLoops()
+	if len(m.loops) != 0 {
+		t.Fatalf("stopAllLoops should clear all loops")
+	}
+}
+
+func TestFireDueLoopSkipsWhenBusy(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "x", time.Minute)
+	m.pending = true // a turn is in flight
+	got, cmd := m.fireDueLoopIfIdle()
+	if got.activeLoopID != "" || cmd != nil {
+		t.Fatalf("a due loop must not fire while a turn is pending")
+	}
+}
+
+func TestFireDueLoopSkipsBehindModal(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "poll", time.Minute) // due immediately
+	m.picker = &commandPicker{}                // e.g. an open /resume session picker
+	got, cmd := m.fireDueLoopIfIdle()
+	if got.activeLoopID != "" || cmd != nil {
+		t.Fatal("a due loop must not launch a run behind an open picker/modal — it would complete into whatever session the user switches to")
+	}
+	if got.loops[0].nextRunAt.IsZero() {
+		t.Fatal("a loop skipped because a modal is open must stay scheduled, not have its next run cleared")
+	}
+}
+
+func TestAdaptiveSelfPaceDelayWidensOverTime(t *testing.T) {
+	if adaptiveSelfPaceDelay(0) >= adaptiveSelfPaceDelay(5) {
+		t.Error("early iterations should check back sooner than later ones")
+	}
+	if adaptiveSelfPaceDelay(20) != 30*time.Minute {
+		t.Errorf("a mature loop should settle to a 30m heartbeat, got %v", adaptiveSelfPaceDelay(20))
+	}
+}
+
+func TestLoopFooterSummary(t *testing.T) {
+	now := time.Date(2026, 7, 5, 15, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	if m.loopFooterSummary() != "" {
+		t.Error("no loops -> empty footer summary")
+	}
+	m = startFixedLoop(m, "a", 5*time.Minute)
+	m.loops[0].nextRunAt = now.Add(5 * time.Minute)
+	got := m.loopFooterSummary()
+	wantWake := "next " + now.Add(5*time.Minute).Format("3:04pm")
+	if !strings.HasPrefix(got, "1 loop") || !strings.Contains(got, wantWake) {
+		t.Fatalf("footer summary = %q, want it to start with a loop count and contain %q", got, wantWake)
+	}
+}
+
+func TestLoopCommandThroughSubmit(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m.input.SetValue("/loop 5m tidy the imports")
+	updated, _ := m.Update(testKey(tea.KeyEnter))
+	next := updated.(model)
+	if len(next.loops) != 1 {
+		t.Fatalf("submitting /loop should create a loop, got %d", len(next.loops))
+	}
+	if next.loops[0].mode != loopModeFixed || next.loops[0].interval != 5*time.Minute {
+		t.Fatalf("unexpected loop from submit: %+v", next.loops[0])
+	}
+	if !next.loopTicking {
+		t.Fatalf("submitting /loop should start the poll ticker")
+	}
+
+	// /loop stop all through the same path clears it.
+	next.input.SetValue("/loop stop all")
+	updated2, _ := next.Update(testKey(tea.KeyEnter))
+	final := updated2.(model)
+	if len(final.loops) != 0 {
+		t.Fatalf("/loop stop all should clear loops, got %d", len(final.loops))
+	}
+}
+
+func TestStaleLoopTickIgnored(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "x", time.Minute)
+	m.loopSeq = 7
+	// A tick from before the seq changed must be dropped and stop the ticker.
+	updated, cmd := m.Update(loopTickMsg{seq: 3})
+	next := updated.(model)
+	if next.loopTicking {
+		t.Error("a stale loop tick should stop the ticker")
+	}
+	if cmd != nil {
+		t.Error("a stale loop tick should not reschedule")
+	}
+}
+
+func TestLoopStopsAfterConsecutiveFailures(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "flaky", time.Minute)
+	id := m.loops[0].id
+	boom := errors.New("boom")
+	for i := 0; i < loopMaxFailures && len(m.loops) > 0; i++ {
+		m = m.advanceLoop(id, "", boom)
+	}
+	if len(m.loops) != 0 {
+		t.Fatalf("loop should stop after %d consecutive failed iterations", loopMaxFailures)
+	}
+}
+
+func TestLoopFailureCounterResetsOnSuccess(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "recovers", time.Minute)
+	id := m.loops[0].id
+	m = m.advanceLoop(id, "", errors.New("x"))
+	m = m.advanceLoop(id, "", errors.New("x"))
+	m = m.advanceLoop(id, "back to healthy", nil) // a success resets the streak
+	if len(m.loops) != 1 {
+		t.Fatalf("a recovered failure streak should not stop the loop")
+	}
+	if m.loops[0].failRun != 0 {
+		t.Fatalf("a successful iteration should reset failRun, got %d", m.loops[0].failRun)
+	}
+}
+
+func TestCancelRunClearsActiveLoopAndRearms(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "x", 5*time.Minute)
+	m.activeLoopID = m.loops[0].id
+	m.loops[0].nextRunAt = time.Time{} // as if the iteration is running
+	m.pending = true
+	m.activeRunID = 7
+	m.runCancel = func() {}
+
+	m.cancelRun()
+
+	if m.activeLoopID != "" {
+		t.Fatal("cancel should clear the active-loop tag so the next turn isn't misattributed")
+	}
+	if m.loops[0].nextRunAt.IsZero() {
+		t.Fatal("cancel should re-arm the interrupted loop, not leave it running forever")
+	}
+	if !m.loops[0].nextRunAt.Equal(now.Add(5 * time.Minute)) {
+		t.Fatalf("re-armed nextRunAt = %v, want +5m", m.loops[0].nextRunAt)
+	}
+}
+
+func TestValidateLoopTargetRejectsBuiltin(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	if _, ok := m.validateLoopTarget("/model list"); ok {
+		t.Error("looping a built-in command should be rejected")
+	}
+	if _, ok := m.validateLoopTarget("keep improving docs"); !ok {
+		t.Error("a plain prompt should be a valid loop target")
+	}
+	if _, ok := m.validateLoopTarget(""); ok {
+		t.Error("an empty target should be rejected")
+	}
+	// A custom command that does not exist is rejected up front, so it is never
+	// scheduled and then re-sent as literal "/…" text on every tick.
+	if _, ok := m.validateLoopTarget("/babysit-prs"); ok {
+		t.Error("an unresolved custom command should be rejected")
+	}
+	// A registered custom command is accepted.
+	m.userCommands = []usercommands.Command{{Name: "babysit-prs", Template: "check the PRs"}}
+	if _, ok := m.validateLoopTarget("/babysit-prs"); !ok {
+		t.Error("an existing custom command should be a valid loop target")
+	}
+}
+
+func TestSelfPacedCommandLoopRejected(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m.userCommands = []usercommands.Command{{Name: "babysit-prs", Template: "check the PRs"}}
+	// A /command with no interval parses as self-paced, but self-paced mode drives
+	// a free-form goal via the playbook + LOOP: protocol, which a command's
+	// expanded template never carries — so a bare self-paced /command is rejected.
+	m, _ = m.startLoop(loopCommand{action: loopActionStart, mode: loopModeSelfPaced, prompt: "/babysit-prs"})
+	if len(m.loops) != 0 {
+		t.Fatal("a bare self-paced /command loop should be rejected (a command loop needs an interval)")
+	}
+	// The same command with an interval (fixed) is accepted.
+	m, _ = m.startLoop(loopCommand{action: loopActionStart, mode: loopModeFixed, prompt: "/babysit-prs", interval: 5 * time.Minute})
+	if len(m.loops) != 1 {
+		t.Fatal("a fixed-interval /command loop should be accepted")
+	}
+}
+
+func TestClearLoopsForSessionSwitch(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "a", time.Minute)
+	m = startFixedLoop(m, "b", time.Minute)
+	m.activeLoopID = m.loops[0].id
+	m, n := m.clearLoopsForSessionSwitch()
+	if n != 2 || len(m.loops) != 0 || m.activeLoopID != "" || m.loopTicking {
+		t.Fatalf("session switch should clear all loop state; got n=%d loops=%d active=%q ticking=%v",
+			n, len(m.loops), m.activeLoopID, m.loopTicking)
+	}
+}
+
+func TestStartNewSessionStopsLoops(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "x", time.Minute)
+	m = m.startNewSession()
+	if len(m.loops) != 0 || m.loopTicking {
+		t.Fatal("/new should stop loops tied to the previous session")
+	}
+}
+
+func TestLoopClearRequiresConfirm(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "x", time.Minute)
+
+	// First /clear with a loop active arms the confirm and does not clear yet.
+	m.input.SetValue("/clear")
+	u1, _ := m.Update(testKey(tea.KeyEnter))
+	n1 := u1.(model)
+	if n1.loopLeavePrompt != commandClear {
+		t.Fatal("first /clear with active loops should arm the confirm, not clear")
+	}
+	if len(n1.loops) != 1 {
+		t.Fatal("the loop should still be active after the warning")
+	}
+
+	// A different command disarms the confirm.
+	n1.input.SetValue("/loop list")
+	u2, _ := n1.Update(testKey(tea.KeyEnter))
+	n2 := u2.(model)
+	if n2.loopLeavePrompt != commandEmpty {
+		t.Fatal("an intervening command should disarm the leave confirm")
+	}
+
+	// Re-arm, then a second consecutive /clear confirms (disarms + clears).
+	n2.input.SetValue("/clear")
+	u3, _ := n2.Update(testKey(tea.KeyEnter))
+	n3 := u3.(model)
+	n3.input.SetValue("/clear")
+	u4, _ := n3.Update(testKey(tea.KeyEnter))
+	n4 := u4.(model)
+	if n4.loopLeavePrompt != commandEmpty {
+		t.Fatal("a second consecutive /clear should confirm and disarm")
+	}
+	if len(n4.loops) != 1 {
+		t.Fatal("/clear keeps loops running (it wipes the screen, not the session)")
+	}
+	// The confirmed clear must actually wipe the transcript, not silently no-op:
+	// only the clear path appends the "Transcript cleared" note, and the pre-clear
+	// loop-start row must be gone.
+	clearedNote, staleRow := false, false
+	for _, r := range n4.transcript {
+		if strings.Contains(r.text, "Transcript cleared") {
+			clearedNote = true
+		}
+		if strings.Contains(r.text, "Loop L1 started") {
+			staleRow = true
+		}
+	}
+	if !clearedNote || staleRow {
+		t.Fatalf("the second /clear must wipe the transcript (clearedNote=%v staleRow=%v)", clearedNote, staleRow)
+	}
+}
+
+func TestLeaveConfirmDisarmedByQueuedPrompt(t *testing.T) {
+	now := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	m := loopTestModel(t, now)
+	m = startFixedLoop(m, "x", time.Minute)
+
+	// Arm the /clear confirm.
+	m.input.SetValue("/clear")
+	u1, _ := m.Update(testKey(tea.KeyEnter))
+	n1 := u1.(model)
+	if n1.loopLeavePrompt != commandClear {
+		t.Fatal("first /clear should arm the confirm")
+	}
+
+	// A prompt submitted while a run streams is queued via an early return in
+	// handleSubmit; it must still disarm the confirm (the disarm runs before those
+	// returns). Otherwise a later /clear would be treated as the second press.
+	n1.pending = true
+	n1.input.SetValue("a follow-up question")
+	u2, _ := n1.Update(testKey(tea.KeyEnter))
+	n2 := u2.(model)
+	if n2.loopLeavePrompt != commandEmpty {
+		t.Fatal("a queued prompt should disarm the leave confirm; the early return must not skip the disarm")
+	}
+}

internal/tui/loop_test.go

@@ -0,0 +1,156 @@
+package tui
+
+import (
+	"strings"
+	"testing"
+	"time"
+)
+
+func TestParseLoopCommandModes(t *testing.T) {
+	cases := []struct {
+		in       string
+		action   loopAction
+		mode     loopMode
+		prompt   string
+		interval time.Duration
+	}{
+		{"", loopActionUsage, 0, "", 0},
+		{"list", loopActionList, 0, "", 0},
+		{"status", loopActionList, 0, "", 0},
+		{"stop all", loopActionStopAll, 0, "", 0},
+		{"stop ab12", loopActionStop, 0, "", 0},
+		{"5m /babysit-prs", loopActionStart, loopModeFixed, "/babysit-prs", 5 * time.Minute},
+		{"90s check the build", loopActionStart, loopModeFixed, "check the build", 90 * time.Second},
+		{"watch CI every 2m", loopActionStart, loopModeFixed, "watch CI", 2 * time.Minute},
+		{"keep improving the docs", loopActionStart, loopModeSelfPaced, "keep improving the docs", 0},
+		// "every PR" is not a cadence — the token after "every" isn't a duration.
+		{"review every PR", loopActionStart, loopModeSelfPaced, "review every PR", 0},
+	}
+	for _, c := range cases {
+		got := parseLoopCommand(c.in)
+		if got.action != c.action {
+			t.Errorf("%q action = %v, want %v", c.in, got.action, c.action)
+		}
+		if c.action == loopActionStart {
+			if got.mode != c.mode {
+				t.Errorf("%q mode = %v, want %v", c.in, got.mode, c.mode)
+			}
+			if got.prompt != c.prompt {
+				t.Errorf("%q prompt = %q, want %q", c.in, got.prompt, c.prompt)
+			}
+			if got.interval != c.interval {
+				t.Errorf("%q interval = %v, want %v", c.in, got.interval, c.interval)
+			}
+		}
+	}
+}
+
+func TestParseLoopCommandStopTarget(t *testing.T) {
+	if got := parseLoopCommand("stop"); got.action != loopActionStop || got.targetID != "" {
+		t.Fatalf("bare stop = %+v, want stop with empty target", got)
+	}
+	if got := parseLoopCommand("stop XyZ9"); got.targetID != "XyZ9" {
+		t.Fatalf("stop target = %q, want XyZ9", got.targetID)
+	}
+}
+
+func TestClampLoopInterval(t *testing.T) {
+	if d, note := clampLoopInterval(5 * time.Second); d != loopMinInterval || note == "" {
+		t.Fatalf("below-min = (%v,%q), want raised to min with note", d, note)
+	}
+	if d, note := clampLoopInterval(48 * time.Hour); d != loopMaxInterval || note == "" {
+		t.Fatalf("above-max = (%v,%q), want lowered to max with note", d, note)
+	}
+	if d, note := clampLoopInterval(5 * time.Minute); d != 5*time.Minute || note != "" {
+		t.Fatalf("in-range = (%v,%q), want unchanged, no note", d, note)
+	}
+}
+
+func TestClampSelfPaceDelay(t *testing.T) {
+	if got := clampSelfPaceDelay(0); got != loopSelfPaceDefault {
+		t.Errorf("zero delay = %v, want default %v", got, loopSelfPaceDefault)
+	}
+	if got := clampSelfPaceDelay(10 * time.Second); got != loopSelfPaceMin {
+		t.Errorf("below-min = %v, want %v", got, loopSelfPaceMin)
+	}
+	if got := clampSelfPaceDelay(3 * time.Hour); got != loopSelfPaceMax {
+		t.Errorf("above-max = %v, want %v", got, loopSelfPaceMax)
+	}
+	if got := clampSelfPaceDelay(10 * time.Minute); got != 10*time.Minute {
+		t.Errorf("in-band = %v, want unchanged", got)
+	}
+}
+
+func TestFormatLoopDuration(t *testing.T) {
+	cases := map[time.Duration]string{
+		30 * time.Second: "30s",
+		90 * time.Second: "90s",
+		5 * time.Minute:  "5m",
+		2 * time.Hour:    "2h",
+		24 * time.Hour:   "1d",
+	}
+	for d, want := range cases {
+		if got := formatLoopDuration(d); got != want {
+			t.Errorf("formatLoopDuration(%v) = %q, want %q", d, got, want)
+		}
+	}
+}
+
+func TestParseLoopControl(t *testing.T) {
+	cases := []struct {
+		in   string
+		done bool
+		wake time.Duration
+	}{
+		{"did the work\nLOOP: DONE", true, 0},
+		{"progress\nLOOP: CONTINUE", false, 0},
+		{"progress\nLOOP: CONTINUE 10m", false, 10 * time.Minute},
+		{"progress\n- loop: continue 1h", false, time.Hour},                     // markdown bullet, lowercase
+		{"LOOP: CONTINUE 5m\n...more...\nLOOP: DONE", true, 0},                  // last control line wins
+		{"no control line here", false, 0},                                      // absent -> adaptive
+		{"LOOP: DONE — all tests pass", true, 0},                                // a trailing note is fine
+		{"progress\nLOOP: CONTINUE 5m — still tidying", false, 5 * time.Minute}, // note after the interval
+		{"work done.\r\nLOOP: DONE\r\n", true, 0},                               // CRLF endings still parse
+		{"work\r\nLOOP: CONTINUE 10m\r\n", false, 10 * time.Minute},             // CRLF + interval
+		{"I will emit LOOP: DONE when finished", false, 0},                      // mid-sentence mention doesn't match
+		{"LOOP: CONTINUE 99z", false, 0},                                        // bad unit -> continue, no wake hint
+	}
+	for _, c := range cases {
+		done, wake := parseLoopControl(c.in)
+		if done != c.done || wake != c.wake {
+			t.Errorf("parseLoopControl(%q) = (%v,%v), want (%v,%v)", c.in, done, wake, c.done, c.wake)
+		}
+	}
+}
+
+func TestFirePromptWrapsSelfPacedOnly(t *testing.T) {
+	self := &loopState{mode: loopModeSelfPaced, prompt: "tidy the docs"}
+	got := self.firePrompt()
+	if !strings.Contains(got, "LOOP: DONE") || !strings.Contains(got, "tidy the docs") {
+		t.Errorf("self-paced firePrompt should embed the playbook and goal, got %q", got)
+	}
+	fixed := &loopState{mode: loopModeFixed, prompt: "check the build"}
+	if fixed.firePrompt() != "check the build" {
+		t.Errorf("fixed-interval firePrompt should be verbatim, got %q", fixed.firePrompt())
+	}
+	// A self-paced /command is a real command, not a goal — send it verbatim, no playbook.
+	cmd := &loopState{mode: loopModeSelfPaced, prompt: "/babysit"}
+	if cmd.firePrompt() != "/babysit" {
+		t.Errorf("self-paced /command firePrompt should be verbatim, got %q", cmd.firePrompt())
+	}
+}
+
+func TestLoopStateDue(t *testing.T) {
+	base := time.Date(2026, 7, 5, 12, 0, 0, 0, time.UTC)
+	l := &loopState{nextRunAt: base}
+	if l.due(base.Add(-time.Second)) {
+		t.Error("loop should not be due before nextRunAt")
+	}
+	if !l.due(base) {
+		t.Error("loop should be due at nextRunAt")
+	}
+	l.paused = true
+	if l.due(base.Add(time.Hour)) {
+		t.Error("paused loop should never be due")
+	}
+}

internal/tui/model.go

@@ -29,6 +29,7 @@ import (
 	"github.com/Gitlawb/zero/internal/providers/providerio"
 	"github.com/Gitlawb/zero/internal/sandbox"
 	"github.com/Gitlawb/zero/internal/sessions"
+	"github.com/Gitlawb/zero/internal/skills"
 	"github.com/Gitlawb/zero/internal/streamjson"
 	"github.com/Gitlawb/zero/internal/tools"
 	"github.com/Gitlawb/zero/internal/usage"
@@ -62,6 +63,7 @@ type model struct {
 	ctx                         context.Context
 	cwd                         string
 	userCommands                []usercommands.Command // file-sourced /commands (.zero/commands)
+	loadSkills                  func() []skills.Skill  // lazy installed-skills loader for /skills + /<skill-name>
 	userConfigPath              string
 	doctorUserConfigPath        string
 	projectConfigPath           string
@@ -188,10 +190,22 @@ type model struct {
 	// like a frozen terminal (for ANY provider, not just slow ones). Zero = idle.
 	turnStartedAt time.Time
 	queuedMessage string
-	exiting       bool
-	runCancel     context.CancelFunc
-	runID         int
-	activeRunID   int
+	// loops holds the session's active /loop definitions (see loop.go). activeLoopID
+	// tags the in-flight run when it is a loop iteration (empty = a user turn), so the
+	// completion seam knows whether to advance a loop. loopSeq invalidates a stale
+	// pending poll tick when loops are stopped; loopTicking guards against scheduling
+	// a second poll ticker. loopLeavePrompt arms a one-shot confirm before /clear or
+	// /quit while loops are active (see handleSubmit).
+	loops           []*loopState
+	activeLoopID    string
+	loopSeq         int
+	loopCounter     int
+	loopTicking     bool
+	loopLeavePrompt commandKind
+	exiting         bool
+	runCancel       context.CancelFunc
+	runID           int
+	activeRunID     int
 	// flushRunIDs holds the ids of runs cancelled while still in flight, mapped
 	// to the session they were recording into AT CANCEL TIME. Each cancelled
 	// agent goroutine keeps running to completion and returns its accumulated
@@ -735,6 +749,7 @@ func newModel(ctx context.Context, options Options) model {
 		cwd:                         cwd,
 		swarmDoneAt:                 map[string]time.Time{},
 		userCommands:                loadedUserCommands,
+		loadSkills:                  options.LoadSkills,
 		composerCursorVisible:       true,
 		userConfigPath:              options.UserConfigPath,
 		doctorUserConfigPath:        doctorUserConfigPath,
@@ -1354,7 +1369,7 @@ func (m model) updateModel(msg tea.Msg) (tea.Model, tea.Cmd) {
 			// home screen or a narrow terminal. Hiding reflows the chat to full
 			// width, so mirror the width-change bookkeeping (re-wrap the streaming
 			// fade, resize the composer) the WindowSizeMsg path does.
-			if m.noBlockingModal() && m.sidebarToggleAllowed() {
+			if !m.transcriptDetailed && m.noBlockingModal() && m.sidebarToggleAllowed() {
 				// Just show/hide — no transcript notice. The reflow IS the feedback,
 				// and emitting a line every toggle piled up noise in the chat.
 				m.sidebarHidden = !m.sidebarHidden
@@ -1421,22 +1436,50 @@ func (m model) updateModel(msg tea.Msg) (tea.Model, tea.Cmd) {
 				return m, nil
 			}
 		case keyIs(msg, tea.KeyPgUp):
-			if m.transcriptDetailed {
-				return m, nil
-			}
-			// A stationary mouse over a bodyY-keyed transcript hover target would
-			// otherwise stay lit at the scrolled-away bodyY until the next real
-			// motion event (see clearHover) — same reasoning as the wheel-scroll
-			// cases in mouse.go.
 			m = m.clearHover()
 			return m.scrollChat(m.chatPageScrollLines()), nil
 		case keyIs(msg, tea.KeyPgDown):
+			m = m.clearHover()
+			return m.scrollChat(-m.chatPageScrollLines()), nil
+		case keyShift(msg) && keyIs(msg, tea.KeyUp):
+			// Shift+Up scrolls the transcript up one line. Must be checked before
+			// plain KeyUp so shifted arrows aren't consumed by the composer-path.
 			if m.transcriptDetailed {
 				return m, nil
 			}
+			if m.pendingPermission != nil {
+				return m.movePermissionCursor(-1), nil
+			}
+			if m.pendingAskUser != nil {
+				return m.moveAskUserCursor(-1), nil
+			}
+			if m.providerWizard != nil {
+				return m.handleProviderWizardKey(msg)
+			}
+			if m.mcpAddWizard != nil {
+				return m.handleMCPAddWizardKey(msg)
+			}
+			if m.mcpManager != nil {
+				return m.handleMCPManagerKey(msg)
+			}
+			if m.picker != nil {
+				if m.modelPickerIsLoading() {
+					return m, nil
+				}
+				m.pickerMoved(-1)
+				return m, nil
+			}
+			if m.suggestionsActive() {
+				break
+			}
+			if m.composerValue() != "" {
+				break // let the input handle multiline navigation
+			}
 			m = m.clearHover()
-			return m.scrollChat(-m.chatPageScrollLines()), nil
-		case keyIs(msg, tea.KeyDown):
+			return m.scrollChat(1), nil
+		case keyShift(msg) && keyIs(msg, tea.KeyDown):
+			// Shift+Down scrolls the transcript down one line. Must be checked
+			// before plain KeyDown so shifted arrows aren't consumed.
 			if m.transcriptDetailed {
 				return m, nil
 			}
@@ -1462,6 +1505,41 @@ func (m model) updateModel(msg tea.Msg) (tea.Model, tea.Cmd) {
 				m.pickerMoved(1)
 				return m, nil
 			}
+			if m.suggestionsActive() {
+				break
+			}
+			if m.composerValue() != "" {
+				break // let the input handle multiline navigation
+			}
+			m = m.clearHover()
+			return m.scrollChat(-1), nil
+		case keyIs(msg, tea.KeyDown):
+			if m.transcriptDetailed {
+				m = m.clearHover()
+				return m.scrollChat(-1), nil
+			}
+			if m.pendingPermission != nil {
+				return m.movePermissionCursor(1), nil
+			}
+			if m.pendingAskUser != nil {
+				return m.moveAskUserCursor(1), nil
+			}
+			if m.providerWizard != nil {
+				return m.handleProviderWizardKey(msg)
+			}
+			if m.mcpAddWizard != nil {
+				return m.handleMCPAddWizardKey(msg)
+			}
+			if m.mcpManager != nil {
+				return m.handleMCPManagerKey(msg)
+			}
+			if m.picker != nil {
+				if m.modelPickerIsLoading() {
+					return m, nil
+				}
+				m.pickerMoved(1)
+				return m, nil
+			}
 			if m.suggestionsActive() {
 				m.moveSuggestion(1)
 				return m, nil
@@ -1480,7 +1558,8 @@ func (m model) updateModel(msg tea.Msg) (tea.Model, tea.Cmd) {
 				return m, nil
 			}
 			if m.transcriptDetailed {
-				return m, nil
+				m = m.clearHover()
+				return m.scrollChat(1), nil
 			}
 			if m.pendingPermission != nil {
 				return m.movePermissionCursor(-1), nil
@@ -1514,6 +1593,48 @@ func (m model) updateModel(msg tea.Msg) (tea.Model, tea.Cmd) {
 			if m.historyRecallActive() {
 				return m.recallHistory(-1), nil
 			}
+		case keyCtrl(msg, 'u'):
+			// Ctrl+U scrolls up half a page, or moves the cursor up in
+			// permission/ask-user prompts. Falls through to the active modal
+			// (wizard, etc.) when none of the above are in focus.
+			if m.transcriptDetailed {
+				return m, nil
+			}
+			if m.pendingPermission != nil {
+				return m.movePermissionCursor(-1), nil
+			}
+			if m.pendingAskUser != nil {
+				return m.moveAskUserCursor(-1), nil
+			}
+			if m.providerWizard != nil || m.mcpAddWizard != nil || m.mcpManager != nil || m.picker != nil || m.pendingSpecReview != nil {
+				break
+			}
+			if m.composerValue() != "" {
+				break // let the input handle its own Ctrl+U (delete-to-bol)
+			}
+			m = m.clearHover()
+			return m.scrollChat(m.chatPageScrollLines()), nil
+		case keyCtrl(msg, 'd'):
+			// Ctrl+D scrolls down half a page, or moves the cursor down in
+			// permission/ask-user prompts. Falls through to the active modal
+			// when none of the above are in focus.
+			if m.transcriptDetailed {
+				return m, nil
+			}
+			if m.pendingPermission != nil {
+				return m.movePermissionCursor(1), nil
+			}
+			if m.pendingAskUser != nil {
+				return m.moveAskUserCursor(1), nil
+			}
+			if m.providerWizard != nil || m.mcpAddWizard != nil || m.mcpManager != nil || m.picker != nil || m.pendingSpecReview != nil {
+				break
+			}
+			if m.composerValue() != "" {
+				break // let the input handle its own Ctrl+D (delete-next-char)
+			}
+			m = m.clearHover()
+			return m.scrollChat(-m.chatPageScrollLines()), nil
 		}
 		if m.transcriptDetailed {
 			return m, nil
@@ -1787,6 +1908,16 @@ func (m model) updateModel(msg tea.Msg) (tea.Model, tea.Cmd) {
 		m.clearComposer()
 		m.clearSuggestions()
 		return m, nil
+	case loopTickMsg:
+		// Idle poll for due loops. A stale tick (loops changed) or an empty loop set
+		// ends the ticker; otherwise fire the earliest due loop if idle and reschedule.
+		if msg.seq != m.loopSeq || len(m.loops) == 0 {
+			m.loopTicking = false
+			return m, nil
+		}
+		var fireCmd tea.Cmd
+		m, fireCmd = m.fireDueLoopIfIdle()
+		return m, tea.Batch(fireCmd, m.scheduleLoopTick())
 	case agentResponseMsg:
 		if msg.runID != m.activeRunID {
 			// A run cancelled while in flight still finishes in its goroutine and
@@ -1956,8 +2087,25 @@ func (m model) updateModel(msg tea.Msg) (tea.Model, tea.Cmd) {
 		// complete once the turn settles.
 		var sweepCmd tea.Cmd
 		m, sweepCmd = m.maybeGitSweep()
+		// If this run was a loop iteration, advance that loop (schedule its next wake
+		// or stop it). Done before launchQueuedMessageIfReady so a user's queued prompt
+		// still wins the immediate re-launch; the loop fires on the next idle tick.
+		var loopTickCmd tea.Cmd
+		if loopID := m.activeLoopID; loopID != "" {
+			m.activeLoopID = ""
+			loopFinalAnswer := ""
+			for _, row := range msg.rows {
+				if row.kind == rowAssistant && row.final {
+					loopFinalAnswer = row.text
+				}
+			}
+			m = m.advanceLoop(loopID, loopFinalAnswer, msg.err)
+			// advanceLoop -> removeLoop may have stopped the ticker; restart it if
+			// other loops remain.
+			m, loopTickCmd = m.ensureLoopTick()
+		}
 		next, queuedCmd := m.launchQueuedMessageIfReady()
-		return next, tea.Batch(titleCmd, recapCmd, sweepCmd, queuedCmd)
+		return next, tea.Batch(titleCmd, recapCmd, sweepCmd, queuedCmd, loopTickCmd)
 	case sessionTitleGeneratedMsg:
 		return m.handleSessionTitleGenerated(msg)
 	case recapGeneratedMsg:
@@ -2215,14 +2363,23 @@ func (m model) View() tea.View {
 	}
 	view.ReportFocus = m.notifier != nil
 	if m.wantsMouseCapture() {
-		// AllMotion (not CellMotion) is required for hover highlighting: it
-		// reports cursor movement even with no button pressed. CellMotion only
-		// reports motion while a button is held (drag) — see bubbletea's
-		// MouseMode docs. AllMotion has marginally worse terminal compatibility
-		// but is well supported by the terminals this app targets; the existing
-		// 15ms mouse-event throttle (mouseEventThrottleInterval) already bounds
-		// the redraw rate from the extra motion events.
-		view.MouseMode = tea.MouseModeAllMotion
+		if isRunningUnderPRoot() {
+			// Under PRoot the AllMotion (1003) sequence doesn't work
+			// reliably, breaking touch-gesture scrolling. Fall back to
+			// CellMotion which still delivers wheel events, clicks, and
+			// drag — the only thing lost is hover-highlighting.
+			view.MouseMode = tea.MouseModeCellMotion
+		} else {
+			// AllMotion (not CellMotion) is required for hover highlighting:
+			// it reports cursor movement even with no button pressed.
+			// CellMotion only reports motion while a button is held (drag) —
+			// see bubbletea's MouseMode docs. AllMotion has marginally worse
+			// terminal compatibility but is well supported by the terminals
+			// this app targets; the existing 15ms mouse-event throttle
+			// (mouseEventThrottleInterval) already bounds the redraw rate
+			// from the extra motion events.
+			view.MouseMode = tea.MouseModeAllMotion
+		}
 	}
 	return view
 }
@@ -2297,7 +2454,7 @@ func (m model) transcriptView() string {
 	if m.transcriptEmpty() && !m.pending && viewportOverlay != "" {
 		emptyOverlay = viewportOverlay
 	}
-	bodyItems := m.transcriptBodyItems(width, emptyOverlay)
+	bodyItems := m.transcriptBodyItems(width, emptyOverlay, false)
 
 	footer := m.footerView(width)
 
@@ -2337,7 +2494,7 @@ func (m model) twoColumnTranscriptView() string {
 	width := chatW
 
 	suggestionOverlay := m.suggestionOverlay(width)
-	bodyItems := m.transcriptBodyItems(width, "")
+	bodyItems := m.transcriptBodyItems(width, "", false)
 	footer := m.footerView(width)
 	overlayForViewport := suggestionOverlay
 	if m.transcriptEmpty() && !m.pending {
@@ -2767,7 +2924,15 @@ func (m model) chatTranscriptViewport() (transcriptViewport, bool) {
 		return transcriptViewport{}, false
 	}
 	width := m.chatColumnWidth()
-	items := m.transcriptBodyItems(width, "")
+	if m.transcriptDetailed {
+		items := m.transcriptBodyItems(width, "", true)
+		body := measureTranscriptBodyItems(items, m.transcriptBodyHeights)
+		header := detailedTranscriptHeader(width) + "\n" + zeroTheme.line.Render(strings.Repeat("-", width))
+		footer := m.detailedTranscriptFooter(width)
+		frame := m.scrollableTranscriptFrame(header, footer)
+		return transcriptViewportForLayout(body, frame, m.chatScrollOffset), true
+	}
+	items := m.transcriptBodyItems(width, "", false)
 	body := measureTranscriptBodyItems(items, m.transcriptBodyHeights)
 	frame := m.scrollableTranscriptFrame(m.pinnedTitleBar(width), m.footerView(width))
 	return transcriptViewportForLayout(body, frame, m.chatScrollOffset), true
@@ -3607,6 +3772,11 @@ func (m model) choosePicker() (tea.Model, tea.Cmd) {
 		if text != "" {
 			m.transcript = reduceTranscript(m.transcript, transcriptAction{kind: actionAppendSystem, text: text})
 		}
+	case pickerSkill:
+		// Fill the composer with "/name " so the user adds their request before
+		// submitting (a bare second Enter runs it without one); names the slash
+		// path cannot reach run immediately instead.
+		m, cmd = m.chooseSkillFromPicker(item)
 	case pickerTheme:
 		// The hovered palette is already live from the preview; handleThemeCommand
 		// records the choice (m.themeMode) and re-applies it, and reports the switch.
@@ -3658,6 +3828,13 @@ func (m model) handleSubmit() (tea.Model, tea.Cmd) {
 		return m, nil
 	}
 	command := parseCommand(input)
+	// A pending /clear or /quit leave-confirmation is armed for the immediately-next
+	// repeat of that same command only; any other submission — including one queued
+	// or deferred by the early returns below — disarms it. Runs before those returns
+	// so an interposed prompt can't be skipped over and leave the gate falsely armed.
+	if m.loopLeavePrompt != commandEmpty && command.kind != m.loopLeavePrompt {
+		m.loopLeavePrompt = commandEmpty
+	}
 	// While exiting (Ctrl+C waiting on the cancelled run's checkpoint flush) a
 	// new run must not start: the deferred tea.Quit would abort it mid-flight
 	// and orphan its checkpoint blobs — the exact loss flushRunIDs prevents.
@@ -3691,6 +3868,14 @@ func (m model) handleSubmit() (tea.Model, tea.Cmd) {
 		m.transcript = reduceTranscript(m.transcript, transcriptAction{kind: actionAppendSystem, text: helpText()})
 		return m, nil
 	case commandClear:
+		// A foreground loop keeps firing after /clear (it wipes the screen, not the
+		// session), so warn once before clearing the context it will run into.
+		if m.loopActive() && m.loopLeavePrompt != commandClear {
+			m.loopLeavePrompt = commandClear
+			m = m.appendLoopSystem(m.loopFooterSummary() + " still running — /clear keeps them firing behind a cleared screen. Run /clear again to confirm, or /loop stop all first.")
+			return m, nil
+		}
+		m.loopLeavePrompt = commandEmpty
 		m.transcript = reduceTranscript(m.transcript, transcriptAction{kind: actionClear})
 		// Clearing wipes the visible transcript only — the session's context is
 		// intact, so the next prompt still replays the full history. Say so, and
@@ -3699,6 +3884,11 @@ func (m model) handleSubmit() (tea.Model, tea.Cmd) {
 		// Scrollback above can't be un-printed; a faint divider marks where the
 		// cleared surface ended and the frontier restarts for the fresh transcript.
 		m.resetFlushFrontier("· cleared ·")
+		// /clear wipes the transcript but not the session, so loops keep running;
+		// say so rather than let them silently fire into a "cleared" screen.
+		if m.loopActive() {
+			m = m.appendLoopSystem(m.loopFooterSummary() + " still running — /loop stop all to end them.")
+		}
 		return m, nil
 	case commandNew:
 		// A fresh session mid-run would strand the in-flight turn's events; make the
@@ -3709,7 +3899,17 @@ func (m model) handleSubmit() (tea.Model, tea.Cmd) {
 			return m, nil
 		}
 		return m.startNewSession(), nil
+	case commandLoop:
+		return m.handleLoopCommand(command.text)
 	case commandExit:
+		// Closing the session stops its foreground loops mid-task; warn once so a
+		// token-spending loop isn't ended by reflex.
+		if m.loopActive() && m.loopLeavePrompt != commandExit {
+			m.loopLeavePrompt = commandExit
+			m = m.appendLoopSystem(m.loopFooterSummary() + " active — closing the session stops them. Run /exit again to confirm, or /loop stop all first.")
+			return m, nil
+		}
+		m.loopLeavePrompt = commandEmpty
 		// /exit gets the same protection as Ctrl+C: cancel any in-flight run and
 		// defer the quit until its checkpoint session events flush — quitting
 		// immediately would orphan the blobs and break /rewind.
@@ -3722,6 +3922,15 @@ func (m model) handleSubmit() (tea.Model, tea.Cmd) {
 	case commandTools:
 		m.transcript = reduceTranscript(m.transcript, transcriptAction{kind: actionAppendSystem, text: m.toolsText()})
 		return m, nil
+	case commandSkills:
+		// With skills installed, /skills opens a searchable picker (like /model);
+		// the text card remains only as the no-skills install hint.
+		if picker := m.newSkillPicker(); picker != nil {
+			m.picker = picker
+			return m, nil
+		}
+		m.transcript = reduceTranscript(m.transcript, transcriptAction{kind: actionAppendSystem, text: m.skillsText()})
+		return m, nil
 	case commandMCP:
 		if strings.TrimSpace(command.text) == "" {
 			return m.openMCPManager(), nil
@@ -3917,6 +4126,11 @@ func (m model) handleSubmit() (tea.Model, tea.Cmd) {
 		if next, cmd, handled := m.handleUserCommand(command.text); handled {
 			return next, cmd
 		}
+		// Then an installed skill: "/<skill-name> [args]" runs the skill directly
+		// (deterministic invocation, vs waiting for the model to pull it in).
+		if next, cmd, handled := m.handleSkillCommand(command.text); handled {
+			return next, cmd
+		}
 		m.transcript = reduceTranscript(m.transcript, transcriptAction{
 			kind: actionAppendError,
 			text: "unknown command: " + command.text,
@@ -4188,6 +4402,21 @@ func (m *model) cancelRun() {
 		m.runCancel()
 	}
 	m.clearStreamingToolCall() // a cancelled file-write must not linger into the next run
+	// A cancelled loop iteration bypasses the agentResponseMsg completion seam (its
+	// late message is drained through flushRunIDs, not advanceLoop), so clear the
+	// loop tag here and re-arm the interrupted loop for its next cadence. Otherwise
+	// the loop is left "running" forever (nextRunAt stays zero) and the NEXT
+	// unrelated turn would be misattributed as this loop's completion.
+	if m.activeLoopID != "" {
+		if l := m.findLoop(m.activeLoopID); l != nil {
+			if l.mode == loopModeSelfPaced {
+				l.nextRunAt = m.now().Add(clampSelfPaceDelay(adaptiveSelfPaceDelay(l.iteration)))
+			} else {
+				l.nextRunAt = m.now().Add(l.interval)
+			}
+		}
+		m.activeLoopID = ""
+	}
 	// Remember the in-flight run — and the session it was recording into — so
 	// its final agentResponseMsg is still drained for session-event persistence
 	// after activeRunID is cleared. Otherwise the checkpoint blobs it captured
@@ -4323,6 +4552,10 @@ func (m model) runAgentWithOptions(runID int, runCtx context.Context, prompt str
 				IncludeLSP:   true,
 				Autonomy:     selfCorrectAutonomyForMode(options.PermissionMode),
 			})
+			// Background post-edit diagnostics: the loop checks files changed by
+			// edit_file/write_file off the tool-call path and nudges the model with
+			// any errors before its next request. Shares the run's lazy manager.
+			options.FileDiagnostics = agent.NewFileDiagnostics(lspManager, options.Cwd)
 		}
 
 		// Some providers synthesize tool-call ids that repeat within a run (e.g.

internal/tui/model_test.go

@@ -2142,7 +2142,7 @@ func TestTranscriptSelectionPaintsHighlightOnceNotTwice(t *testing.T) {
 	width := m.chatColumnWidth()
 
 	findRow := func(mm model) (transcriptBodyItem, bool) {
-		for _, it := range mm.transcriptBodyItems(width, "") {
+		for _, it := range mm.transcriptBodyItems(width, "", false) {
 			if it.kind == transcriptBodyItemRow && it.rowIndex >= 0 {
 				return it, true
 			}
@@ -2198,7 +2198,7 @@ func TestReasoningAfterToolCardGetsBlankSeparator(t *testing.T) {
 		transcriptRow{kind: rowReasoning, text: "Considering the next step in detail before acting"},
 		transcriptRow{kind: rowToolResult, id: "t2", tool: "read_file", status: tools.StatusOK, detail: "x\ny"},
 	)
-	items := m.transcriptBodyItems(m.chatColumnWidth(), "")
+	items := m.transcriptBodyItems(m.chatColumnWidth(), "", false)
 	reasoningIdx := -1
 	for i := range items {
 		if items[i].rowIndex >= 0 && items[i].rowIndex < len(m.transcript) &&
@@ -2223,7 +2223,7 @@ func TestAssistantNarrationBeforeToolCardGetsBlankSeparator(t *testing.T) {
 		transcriptRow{kind: rowAssistant, text: "I'll inspect the existing file, then run it."},
 		transcriptRow{kind: rowToolResult, id: "t1", tool: "read_file", status: tools.StatusOK, detail: "File: time_test.py\n\n  1 | print('x')"},
 	)
-	items := m.transcriptBodyItems(m.chatColumnWidth(), "")
+	items := m.transcriptBodyItems(m.chatColumnWidth(), "", false)
 	toolIdx := -1
 	for i := range items {
 		if items[i].rowIndex >= 0 && items[i].rowIndex < len(m.transcript) &&
@@ -2247,7 +2247,7 @@ func TestUserPromptBeforeToolCardGetsBlankSeparator(t *testing.T) {
 		transcriptRow{kind: rowUser, text: "create the landing page"},
 		transcriptRow{kind: rowToolResult, id: "t1", tool: "write_file", status: tools.StatusOK, detail: "--- /dev/null\n+++ b/index.html\n@@ -0,0 +1,1 @@\n+<!DOCTYPE html>"},
 	)
-	items := m.transcriptBodyItems(m.chatColumnWidth(), "")
+	items := m.transcriptBodyItems(m.chatColumnWidth(), "", false)
 	toolIdx := -1
 	for i := range items {
 		if items[i].rowIndex >= 0 && items[i].rowIndex < len(m.transcript) &&
@@ -2273,7 +2273,7 @@ func TestAssistantAfterToolCardGetsRuleSeparator(t *testing.T) {
 		transcriptRow{kind: rowToolResult, id: "t1", tool: "bash", status: tools.StatusOK, detail: "stdout:\nok\nexit_code: 0"},
 		transcriptRow{kind: rowAssistant, text: "Done.", final: true},
 	)
-	items := m.transcriptBodyItems(m.chatColumnWidth(), "")
+	items := m.transcriptBodyItems(m.chatColumnWidth(), "", false)
 	finalIdx := -1
 	for i := range items {
 		if items[i].rowIndex >= 0 && items[i].rowIndex < len(m.transcript) &&
@@ -2307,7 +2307,7 @@ func TestStreamingAssistantAfterToolCardGetsRuleSeparator(t *testing.T) {
 		transcriptRow{kind: rowAssistant, text: "I'll run it first."},
 		transcriptRow{kind: rowToolResult, id: "t1", tool: "bash", status: tools.StatusOK, detail: "stdout:\nok\nexit_code: 0"},
 	)
-	items := m.transcriptBodyItems(m.chatColumnWidth(), "")
+	items := m.transcriptBodyItems(m.chatColumnWidth(), "", false)
 	pendingIdx := -1
 	for i := range items {
 		if items[i].kind == transcriptBodyItemPendingInterim {
@@ -2357,7 +2357,7 @@ func TestTranscriptBodyRowsUseFullWidthAndAlignSelection(t *testing.T) {
 		t.Fatalf("expected no body gutter at width %d, got %d", width, gutter)
 	}
 
-	items := m.transcriptBodyItems(width, "")
+	items := m.transcriptBodyItems(width, "", false)
 	var row *transcriptBodyItem
 	for i := range items {
 		if items[i].kind == transcriptBodyItemRow && items[i].rowIndex >= 0 {

internal/tui/mouse.go

@@ -1,6 +1,36 @@
 package tui
 
-import tea "charm.land/bubbletea/v2"
+import (
+	"os"
+	"strings"
+	"sync"
+
+	tea "charm.land/bubbletea/v2"
+)
+
+// parseTracerPid returns true when /proc/self/status contains a non-zero
+// TracerPid. This detects any ptrace tracer (PRoot, gdb, strace, dlv, etc.),
+// not PRoot specifically. That's intentional: under any ptrace tracer the
+// AllMotion (1003) sequence is unreliable, so CellMotion is the safer
+// fallback. The only loss is hover-highlighting, which has no functional
+// impact.
+func parseTracerPid(data []byte) bool {
+	for _, line := range strings.Split(string(data), "\n") {
+		if strings.HasPrefix(line, "TracerPid:") {
+			pid := strings.TrimSpace(line[10:])
+			return pid != "0"
+		}
+	}
+	return false
+}
+
+var isRunningUnderPRoot = sync.OnceValue(func() bool {
+	data, err := os.ReadFile("/proc/self/status")
+	if err != nil {
+		return false
+	}
+	return parseTracerPid(data)
+})
 
 type mouseOverlayHit struct {
 	x int
@@ -257,7 +287,7 @@ func (m model) syncMouseCapture() (model, tea.Cmd) {
 }
 
 func (m model) mouseOverComposer(msg tea.MouseMsg) bool {
-	if !m.altScreen || m.height <= 0 {
+	if !m.altScreen || m.height <= 0 || m.transcriptDetailed {
 		return false
 	}
 	width := m.chatColumnWidth()

internal/tui/mouse_test.go

@@ -920,6 +920,30 @@ func TestMouseCaptureOnlyDuringInteractiveSetupStages(t *testing.T) {
 	}
 }
 
+// TestComposerMouseSelectionBlockedInDetailedMode: composer mouse hit-testing
+// is disabled while the detailed transcript view is active, so wheel and click
+// events in the composer area reach the transcript body instead.
+func TestComposerMouseSelectionBlockedInDetailedMode(t *testing.T) {
+	m := mouseTestModel()
+	m.input.SetValue("some text")
+
+	updated, _ := m.Update(testKeyCtrl('o'))
+	m = updated.(model)
+
+	if !m.transcriptDetailed {
+		t.Fatal("sanity check: Ctrl+O should enter detailed mode")
+	}
+	if !m.composerMouseSelectionBlocked() {
+		t.Fatal("composerMouseSelectionBlocked should be true in detailed mode")
+	}
+
+	width := m.chatColumnWidth()
+	frame := m.scrollableTranscriptFrame(m.pinnedTitleBar(width), m.footerView(width))
+	if m.mouseOverComposer(testMouseWheel(tea.MouseWheelUp, 0, frame.composerRect.y)) {
+		t.Fatal("mouseOverComposer should return false in detailed mode")
+	}
+}
+
 func firstTranscriptTextMouseY(t *testing.T, m model) int {
 	t.Helper()
 	_, y := firstTranscriptTextMousePoint(t, m)
@@ -986,6 +1010,38 @@ func withSidebarContent(m model) model {
 	return m
 }
 
+func TestParseTracerPidMissing(t *testing.T) {
+	if got := parseTracerPid([]byte("Pid: 123\nName: foo\n")); got {
+		t.Fatal("missing TracerPid should return false")
+	}
+}
+
+func TestParseTracerPidZero(t *testing.T) {
+	if got := parseTracerPid([]byte("TracerPid: 0\nName: foo\n")); got {
+		t.Fatal("TracerPid: 0 should return false")
+	}
+}
+
+func TestParseTracerPidNonZero(t *testing.T) {
+	if got := parseTracerPid([]byte("Name: foo\nTracerPid: 42\n")); !got {
+		t.Fatal("TracerPid: 42 should return true")
+	}
+}
+
+func TestParseTracerPidMalformed(t *testing.T) {
+	// Non-numeric values are treated as non-zero since any non-"0" string
+	// indicates a tracer is attached, matching the function's semantics.
+	if got := parseTracerPid([]byte("TracerPid: abc\n")); !got {
+		t.Fatal("malformed TracerPid should return true (non-zero string)")
+	}
+}
+
+func TestParseTracerPidEmpty(t *testing.T) {
+	if got := parseTracerPid([]byte{}); got {
+		t.Fatal("empty input should return false")
+	}
+}
+
 func setupMouseTestModel() model {
 	m := newModel(context.Background(), Options{
 		Setup: SetupOptions{

internal/tui/options.go

@@ -13,6 +13,7 @@ import (
 	"github.com/Gitlawb/zero/internal/providermodeldiscovery"
 	"github.com/Gitlawb/zero/internal/sandbox"
 	"github.com/Gitlawb/zero/internal/sessions"
+	"github.com/Gitlawb/zero/internal/skills"
 	"github.com/Gitlawb/zero/internal/tools"
 	"github.com/Gitlawb/zero/internal/usage"
 	"github.com/Gitlawb/zero/internal/zeroruntime"
@@ -48,7 +49,13 @@ type Options struct {
 	SessionCompactor            SessionCompactor
 	PrService                   *PrService
 
-	AgentOptions    agent.Options
+	AgentOptions agent.Options
+	// LoadSkills returns the installed skills (default skills dir merged with any
+	// plugin skill roots), bodies included, for /skills and direct /<skill-name>
+	// invocation. Called lazily per use so newly installed skills are picked up
+	// without a restart. Nil means the session has no skills wiring (skills stay
+	// model-pulled via the skill tool only).
+	LoadSkills      func() []skills.Skill
 	PermissionMode  agent.PermissionMode
 	ReasoningEffort modelregistry.ReasoningEffort
 	ResponseStyle   string

internal/tui/permission_prompt_test.go

@@ -309,3 +309,70 @@ func TestPermissionRenderShowsNetworkTargetAndHostScopedAlways(t *testing.T) {
 		t.Fatalf("network prompt should render target label, got %q", got)
 	}
 }
+
+// TestPermissionCursorCtrlU ensures Ctrl+U moves the permission cursor UP
+// (toward the first option). Regression: the original implementation moved
+// the cursor DOWN on Ctrl+U and UP on Ctrl+D.
+func TestPermissionCursorCtrlU(t *testing.T) {
+	m := pendingPermissionModel(t, func(agent.PermissionDecision) {})
+	m.pendingPermission.cursor = 2 // middle option
+
+	updated, _ := m.Update(testKeyCtrl('u'))
+	next := updated.(model)
+	if next.pendingPermission.cursor != 1 {
+		t.Fatalf("Ctrl+U moved cursor to %d, want 1 (one step up)", next.pendingPermission.cursor)
+	}
+}
+
+// TestPermissionCursorCtrlD ensures Ctrl+D moves the permission cursor DOWN
+// (toward the last option). Regression: the original implementation moved
+// the cursor UP on Ctrl+D and DOWN on Ctrl+U.
+func TestPermissionCursorCtrlD(t *testing.T) {
+	m := pendingPermissionModel(t, func(agent.PermissionDecision) {})
+	m.pendingPermission.cursor = 0 // first option
+
+	updated, _ := m.Update(testKeyCtrl('d'))
+	next := updated.(model)
+	if next.pendingPermission.cursor != 1 {
+		t.Fatalf("Ctrl+D moved cursor to %d, want 1 (one step down)", next.pendingPermission.cursor)
+	}
+}
+
+// TestShiftUpComposerGuard ensures Shift+Up does NOT scroll the transcript
+// when the composer has text, so it falls through to the input's own line
+// navigation.
+func TestShiftUpComposerGuard(t *testing.T) {
+	m := mouseTestModel()
+	// Add enough transcript rows so scrolling is valid.
+	for i := 0; i < 20; i++ {
+		m.transcript = appendRow(m.transcript, rowAssistant, "line")
+	}
+	m.input.SetValue("some text")
+	m.chatScrollOffset = 5
+
+	updated, cmd := m.Update(testKeyShift(tea.KeyUp))
+	next := updated.(model)
+	_ = cmd
+	if got := next.chatScrollOffset; got != 5 {
+		t.Fatalf("Shift+Up with non-empty composer scrolled offset to %d, want 5 (unchanged)", got)
+	}
+}
+
+// TestShiftDownComposerGuard ensures Shift+Down does NOT scroll the transcript
+// when the composer has text, falling through to the input's navigation.
+func TestShiftDownComposerGuard(t *testing.T) {
+	m := mouseTestModel()
+	// Add enough transcript rows so scrolling is valid.
+	for i := 0; i < 20; i++ {
+		m.transcript = appendRow(m.transcript, rowAssistant, "line")
+	}
+	m.input.SetValue("some text")
+	m.chatScrollOffset = 3
+
+	updated, cmd := m.Update(testKeyShift(tea.KeyDown))
+	next := updated.(model)
+	_ = cmd
+	if got := next.chatScrollOffset; got != 3 {
+		t.Fatalf("Shift+Down with non-empty composer scrolled offset to %d, want 3 (unchanged)", got)
+	}
+}

internal/tui/picker.go

@@ -24,6 +24,7 @@ const (
 	pickerEffort
 	pickerSession
 	pickerTheme
+	pickerSkill
 )
 
 // pickerItem is one selectable row: Label is shown, Value is passed to the
@@ -225,6 +226,35 @@ func (m model) newModelPicker() *commandPicker {
 	return &commandPicker{kind: pickerModel, title: "Choose a model", items: items, allItems: append([]pickerItem{}, items...), selected: 0}
 }
 
+// newSkillPicker lists the installed skills for browse-and-run: type to filter,
+// arrow keys to select, Enter runs the chosen skill immediately. Selection is by
+// the skill's EXACT name (item.Value), so even skills the slash path cannot
+// reach — names shadowed by a builtin or user command, or names that don't fit
+// a slash token — are runnable here: picking from the Skills picker is
+// unambiguous. Returns nil when no skills are installed (the caller falls back
+// to the install-hint card).
+func (m model) newSkillPicker() *commandPicker {
+	installed := m.installedSkills()
+	items := make([]pickerItem, 0, len(installed))
+	for _, skill := range installed {
+		name := strings.TrimSpace(skill.Name)
+		if name == "" {
+			continue
+		}
+		label := name
+		if slash := skillSlashName(name); slash != "" {
+			// Show the slash form where one exists — it doubles as documentation
+			// of the direct "/name [args]" invocation.
+			label = "/" + slash
+		}
+		items = append(items, pickerItem{Label: label, Value: name, Meta: strings.TrimSpace(skill.Description)})
+	}
+	if len(items) == 0 {
+		return nil
+	}
+	return &commandPicker{kind: pickerSkill, title: "pick a skill · ⏎ fills /name — add your request", items: items, allItems: append([]pickerItem{}, items...), selected: 0}
+}
+
 // modelPickerProviders returns the providers to list in /model: all saved
 // providers, falling back to the active profile when none were threaded in.
 func (m model) modelPickerProviders() []config.ProviderProfile {

internal/tui/rendering_lime_test.go

@@ -511,6 +511,50 @@ func TestSelectableAssistantRowKeepsMarkdownSemanticsPlain(t *testing.T) {
 	}
 }
 
+func TestStripMarkdownRenderControlsStripsAllANSI(t *testing.T) {
+	// Bold markers that renderAssistantMarkdownText embeds for prose formatting.
+	withBold := "\x1b[1mhello\x1b[22m world"
+	if got := stripMarkdownRenderControls(withBold); got != "hello world" {
+		t.Fatalf("stripMarkdownRenderControls(%q) = %q, want %q", withBold, got, "hello world")
+	}
+	// Color sequences from highlightCodeAuto for fenced code blocks.
+	withColor := "\x1b[38;2;236;236;238mdef\x1b[0m hello():"
+	if got := stripMarkdownRenderControls(withColor); got != "def hello():" {
+		t.Fatalf("stripMarkdownRenderControls(%q) = %q, want %q", withColor, got, "def hello():")
+	}
+	// Sequences that combine both (bold + underline + color).
+	withBoldColor := "\x1b[1;4;38;2;236;236;238mW\x1b[m"
+	if got := stripMarkdownRenderControls(withBoldColor); got != "W" {
+		t.Fatalf("stripMarkdownRenderControls(%q) = %q, want %q", withBoldColor, got, "W")
+	}
+}
+
+func TestSelectedTranscriptTextStripsANSIFromHighlightedCode(t *testing.T) {
+	m := limeTestModel()
+	m.width, m.height = 120, 40
+	m.altScreen = true
+	m.headerPrinted = true
+	code := `package main
+
+import "fmt"
+
+func main() {
+	fmt.Println("hello")
+}`
+	row := transcriptRow{kind: rowAssistant, text: "First, here's the code:\n\n```go\n" + code + "\n```\n\nThat's it.", final: true}
+	m.transcript = append(m.transcript, row)
+	m.flushed = len(m.transcript)
+
+	// The highlighted code block contains ANSI color sequences from chroma;
+	// verify none survive in the selectable text used for clipboard copy.
+	_, selectable := m.renderSelectableAssistantRow(0, row, 72, 0)
+	for _, line := range selectable {
+		if ansiPattern.MatchString(line.text) {
+			t.Fatalf("selectable text leaked ANSI escapes: %q", line.text)
+		}
+	}
+}
+
 func TestFinalAnswerRendersPlainTextAndCompletionLine(t *testing.T) {
 	m := limeTestModel()
 	row := transcriptRow{

internal/tui/session.go

@@ -99,6 +99,12 @@ func (m model) startNewSession() model {
 	// Scrollback above can't be un-printed; a faint divider marks the boundary and
 	// the flush frontier restarts for the fresh transcript (mirrors /clear, /resume).
 	m.resetFlushFrontier("· new session ·")
+	// Loops belong to the previous session; stop them so they don't fire the old
+	// conversation's prompt into the fresh one.
+	if updated, cleared := m.clearLoopsForSessionSwitch(); cleared > 0 {
+		m = updated
+		m.transcript = reduceTranscript(m.transcript, transcriptAction{kind: actionAppendSystem, text: fmt.Sprintf("Stopped %d loop(s) tied to the previous session.", cleared)})
+	}
 	return m
 }
 
@@ -123,13 +129,24 @@ func (m model) appendSessionEvent(eventType sessions.EventType, payload any) (mo
 
 func (m model) appendSessionEvents(events []pendingSessionEvent) (model, []transcriptRow) {
 	rows := []transcriptRow{}
+	if m.activeSession.SessionID == "" || len(events) == 0 {
+		return m, rows
+	}
+	inputs := make([]sessions.AppendEventInput, 0, len(events))
 	for _, event := range events {
-		next, err := m.appendSessionEvent(event.Type, event.Payload)
-		if err != nil {
-			rows = append(rows, transcriptRow{kind: rowError, text: "session record error: " + err.Error()})
-			continue
-		}
-		m = next
+		inputs = append(inputs, sessions.AppendEventInput{Type: event.Type, Payload: event.Payload})
+	}
+	appended, err := m.sessionStore.AppendEvents(m.activeSession.SessionID, inputs)
+	if err != nil {
+		rows = append(rows, transcriptRow{kind: rowError, text: "session record error: " + err.Error()})
+		return m, rows
+	}
+	if len(appended) > 0 {
+		last := appended[len(appended)-1]
+		m.activeSession.UpdatedAt = last.CreatedAt
+		m.activeSession.EventCount = last.Sequence
+		m.activeSession.LastEventType = last.Type
+		m.sessionEvents = append(m.sessionEvents, appended...)
 	}
 	return m, rows
 }
@@ -142,13 +159,12 @@ func (m model) appendSessionEventsTo(sessionID string, events []pendingSessionEv
 	if m.sessionStore == nil || sessionID == "" {
 		return rows
 	}
+	inputs := make([]sessions.AppendEventInput, 0, len(events))
 	for _, event := range events {
-		if _, err := m.sessionStore.AppendEvent(sessionID, sessions.AppendEventInput{
-			Type:    event.Type,
-			Payload: event.Payload,
-		}); err != nil {
-			rows = append(rows, transcriptRow{kind: rowError, text: "session record error: " + err.Error()})
-		}
+		inputs = append(inputs, sessions.AppendEventInput{Type: event.Type, Payload: event.Payload})
+	}
+	if _, err := m.sessionStore.AppendEvents(sessionID, inputs); err != nil {
+		rows = append(rows, transcriptRow{kind: rowError, text: "session record error: " + err.Error()})
 	}
 	return rows
 }
@@ -195,6 +211,10 @@ func (m model) handleResumeCommand(args string) (model, string) {
 		return m, "Sessions\nerror: " + err.Error()
 	}
 
+	// Capture the current session id before switching so loops are only torn down
+	// on a real change — `/resume latest` or `/resume <currentID>` can resolve to
+	// the already-active session, whose loops belong to it, not a "previous" one.
+	previousID := m.activeSession.SessionID
 	m.activeSession = *session
 	m.sessionEvents = append([]sessions.Event{}, events...)
 	if m.providerName == "" {
@@ -203,9 +223,16 @@ func (m model) handleResumeCommand(args string) (model, string) {
 	if m.modelName == "" {
 		m.modelName = session.ModelID
 	}
+	loopsCleared := 0
+	if session.SessionID != previousID {
+		m, loopsCleared = m.clearLoopsForSessionSwitch()
+	}
 
 	rows := initialTranscript()
 	rows = appendRow(rows, rowSystem, m.formatResumeSummary(*session, len(events)))
+	if loopsCleared > 0 {
+		rows = appendRow(rows, rowSystem, fmt.Sprintf("Stopped %d loop(s) tied to the previous session.", loopsCleared))
+	}
 	rows = appendTranscriptRowsDedup(rows, transcriptRowsFromSessionEvents(events))
 	m.transcript = rows
 	// Every rehydrated row is settled by construction, so resetting the flush