Files
zero/internal/tools/tool_search.go
Anandan afb5010a46 polish v2: smoothness & delight (#484)
* feat(tui): actionable hints on provider/model errors

Classify a failed turn's provider error into auth / rate-limit /
connectivity / model-not-found / context-overflow and render a one-line
next step below the red error row, instead of dumping an identical raw
blob for every failure mode.

The classifier lives in a new internal/errhint package so both the TUI
error row and the CLI exec path can share it (TUIHint references slash
commands, CLIHint references zero subcommands). Classification is a
conservative string heuristic keyed off providerio.ClassifiedError's
prefixes plus lower-level DNS/TLS/timeout/context-length signatures,
since the numeric HTTP status is gone by the time the error reaches a UI
surface.

* feat(tools): budget bash output with head+tail truncation

bash was the one tool that returned command output with no byte cap —
'cat large.log' or a verbose test run could dump megabytes into context
and force compaction, while every other read/search tool already applies
a budget.

Cap stdout and stderr at 96KB each, keeping the head and tail of an
oversized stream (build/test failures usually land at the tail) and
dropping the middle behind a marker that points at redirect-to-file +
read_file. Record raw_bytes/emitted_bytes/estimated_tokens/truncated in
Meta like the other tools. Sandbox-denial and shell-issue detection still
run on the full raw output before budgeting.

* fix(gemini): retry 401 with refreshed OAuth token

The Gemini provider used providerio.SendWithRetry (429/503/529 only),
so an expired OAuth token surfaced as a terminal 'API key not valid'
auth error while OpenAI and Anthropic silently recovered via
SendWithAuthRetry's force-refresh-and-replay on 401.

Plumb OAuthResolver through gemini.Options/Provider and the factory
(exactly as the other two providers already do) and switch the stream
send to SendWithAuthRetry. With a nil resolver (API-key users) behavior
is unchanged.

* fix(errhint): gate classification on provider-origin marker

Address CodeRabbit review on #1: agentResponseMsg.err can also carry
local failures (a tool's 'permission denied', a 'file does not exist',
a config error), and broad substrings would have attached a bogus
/provider or /model hint to those.

Classify now returns Unknown unless the message carries a provider marker
the provider layer always attaches (auth error: / rate limit error: /
provider error: / provider request error: / provider stream error:).
Local errors never draw a provider hint; provider errors sub-classify as
before. Added local-failure test cases proving they stay Unknown.

* feat(cli): actionable hint on text-mode exec provider errors

zero exec printed provider failures as a bare '[zero] <raw error>' with
no next step. Append a one-line hint (reusing the shared errhint
classifier) for recognized provider failures — 'run `zero auth`',
'run `zero doctor`', etc. The provider-origin gate means non-provider
codes (sandbox_error, mcp_error) and JSON/stream-json output are
untouched.

* fix(tools): set Result.Truncated for budgeted bash output

Address CodeRabbit review on #3: budgetBashOutput recorded truncation in
meta["truncated"] but never set the Result.Truncated struct field, so
consumers reading the struct (like glob/web_fetch/read_minified do) would
miss truncated bash results. Return the bool from budgetBashOutput and
set Result.Truncated on every return path.

* perf(tui): reuse one LSP manager across prompts

runAgentWithOptions built a fresh lsp.NewManager per run and shut it down
when the run returned, so gopls (and every other language server)
cold-started on the first edit of every turn — 200ms-2s of latency the
manager's own design (long-lived, reused servers) exists to avoid.

Build one manager per session in newModel (cheap: servers start lazily on
first Check) and reuse it across runs; the SelfCorrector still gets a
fresh checker wrapping the shared manager each run. Torn down with a
short deadline in quit(). Runs fall back to a per-run manager when the
session manager is absent (cwd unknown, or a directly-built test model).

* perf(tui): accumulate streamingText as []byte (O(1) append)

m.streamingText += msg.delta was O(len) per delta -> O(n²) across a long
generation (a 10k-token code gen allocates hundreds of MB of intermediate
strings and the UI gets progressively laggier). Accumulate into a []byte
with append (O(1) amortized) and read via streamingTextString().

A []byte rather than strings.Builder because the TUI model is copied by
value on every Update, which would trip strings.Builder's copy check;
[]byte is nil-safe and copies cleanly like the other slice fields.

* perf(tui): coalesce streamed text deltas to one frame

Every OnText delta was its own tea.Msg, so a fast provider (100+ tok/s)
drove 100+ full Update->View cycles per second, each re-parsing the
growing markdown — visible stutter over SSH and wasted CPU.

Batch agentTextMsg deltas at the runtime sink over a ~16ms frame and
forward them as a single message, decoupling render rate from token rate.
Any non-text message flushes pending text first so ordering with
tool-call / reasoning / row messages is preserved; the final
agentResponseMsg (a tea.Cmd return, not a sink message) is safe because
the model already drops deltas for an inactive runID.

* fix(agent): raise tool-failure stop threshold 4->6

guardrails.go halted a run after 4 consecutive same-error tool failures.
A corrective hint fires at 2, and a model iterating on a genuinely tricky
edit can legitimately fail a couple more times after the hint while
converging — stopping at 4 cut those runs short. Raise to 6; the streak
still resets the instant the tool succeeds or hits a different error, so
only true same-error loops are affected.

* fix(agent): more reconnect retries with jitter; narrow to transport errors

#19: bump maxStreamReconnects 2->4 and add up-to-50% jitter on the
exponential backoff (capped at 8s), so a multi-second network blip is
ridden out instead of killing the run on a 2s hiccup, and concurrent runs
(swarms, cron fleets) don't reconnect in lockstep.

#34: drop the '502'/'503' substring matches from shouldReconnect. 503
already exhausted providerio.SendWithRetry (retrying here is a redundant
double-retry) and 502 is non-idempotent by providerio's rule (the POST
may have been processed). Only genuine transport failures — where no
response was received — reconnect now.

* fix(agent): reduce stall retries 2->1 to bound stuck-session time

A no-output stream stall is detected only after the full stream idle
timeout (~5min) elapses, and each retry can idle again — so 2 retries
left an interactive session frozen for ~15min. Drop to 1 retry: keeps the
common single-hiccup recovery while bounding the worst case to ~2x the
idle timeout.

* fix(tui): serialize coalescer forwarding to preserve message order

Address CodeRabbit review on #7: flush() and the run-switch path drained
the text buffer under c.mu but called forward() after releasing it, so a
timer-fired text flush could race a concurrent non-text send() and land
after the tool-call/reasoning message it should precede.

Hold c.mu across drain AND forward (drainAndForwardLocked), so whoever
holds the lock delivers atomically and the other caller blocks until it
is done — text can never overtake a following non-text message. Added a
concurrency stress test (timer racing an inline send) that passes under
-race.

* feat(tui): accept /compact now

Bare /compact already triggers compaction, but /compact now — what users
reach for when the context gauge climbs — hit the usage-error branch.
Accept the 'now' keyword and advertise it in the command usage.

* feat(tui): add /retry, /edit, /copy, /export commands

/retry resends the last prompt; /edit recalls it into the composer to
tweak and resend — both read a new lastPrompt field captured verbatim
(pre-expansion) in launchPrompt. /copy puts the last answer on the
clipboard via the existing copy machinery; /export writes a plain
role-prefixed transcript to a file (timestamped default, or a given
path). Real gaps for SSH users with no mouse select.

* feat(tools): separator-insensitive tool_search matching

tool_search ranked deferred tools by exact substring, so a model that
typed 'webfetch' (dropping the underscore) matched nothing and looped.
Add a separator-squashing fallback ('web_fetch' -> 'webfetch') scored
below an exact substring match, so precise queries still rank first.

* feat(tools): signal ask_user dismissal distinctly from a blank field

Empty answers all rendered as '(no answer provided)', so the model
couldn't tell a wholesale dismissal (user closed the prompt without
answering anything) from one field left blank amid real answers — and
might invent a default. FormatAskUserAnswers now flags a full dismissal
up front as a skip and marks individual empties '(left blank)' vs
'(skipped)'.

* feat(tui): fuzzy ranking in model/session pickers

The pickers filtered by flat strings.Contains with no ranking, so
finding 'sonnet 4.5' among 50+ models meant an exact prefix or scrolling.
Rank matches (exact < prefix < contains < subsequence, reusing
fuzzySubsequenceGap) so the closest match lands on top, and 'snt45' now
matches 'Sonnet 4.5'. Groups stay contiguous — ordered by their best
match, never split — so the grouped model picker still renders one header
per provider. Covers both /model and /resume (shared commandPicker).

* perf(agent): cache per-tool schema render across turns

partitionTools re-ran the recursive schema->map conversion
(schemaToRuntimeMap) for every tool on every turn — redundant, since a
tool's advertised name/description/schema is stable for the run. Add a
per-run definition cache keyed by tool name (nil-safe; the plain
partitionTools entrypoint and tests pass nil for fresh renders).

The partitioning itself (visibility, deferral, ordering) still recomputes
every turn — it must, because a tool's deferred state can flip mid-run
(swarm tools un-defer once a swarm is active). Only the expensive schema
render is memoized. tool_search is excluded by its callers (dynamic
description) so it never poisons the cache.

* perf(agent): calibrate compaction estimate against real prompt tokens

The byte/4 heuristic (ApproxTextTokens) over-counts code-heavy content
~15-20%, so with triggerRatio=0.7 compaction fired at ~60% of true
capacity — premature summarizer calls that degrade quality. Each turn now
folds (rawEstimate, provider InputTokens) into an EMA calibration ratio,
clamped per-sample to a sane band; maybeCompact and the reactive path
scale their estimates by it. Turn 1 uses the raw estimate (no data yet);
later turns compact near real capacity.

* style: gofmt the new agent test files

The CI gofmt check flagged the #22/#25 test files (struct-field and
switch alignment). No behavior change.

* feat(agent): turn-based plan staleness reminder

The stale-plan nudge only fired after 10 tool calls since the last
update_plan, so a plan that drifts stale across many low-tool-call turns
(one call per turn) took many turns to catch — the 'plan set turn 1 stays
pending forever' case. Add a turn-count complement: after
stalePlanTurnThreshold (8) turns without an update AND with items still
pending, fire the same one-shot reminder. Gated on pending items so a
fully-completed plan is never nagged; the existing tool-call trigger is
unchanged. (The roadmap's fuzzy 'tool calls unrelated to plan items'
relevance match is deliberately omitted — too false-positive-prone.)

* feat(agent): preserve recent-edits summary across compaction

Compaction preserved the plan, loaded skills, tool schemas, and project
instructions, but not file edits — so after the editing turns were elided
the model knew it had changed a file only vaguely, and would re-read to
rediscover its own footprint. Extract write_file/edit_file targets from
the elided middle with a one-line note from each tool result, merge them
across repeated compactions (newer note per path wins), and carry them in
the preserved-state JSON (capped at 20 paths / 160 bytes per note).

* feat(tui): workspace-scoped /resume + fewer event reads

/resume listed sessions from every project globally, and read the full
event file for each of them to test for resumable content. Filter the
picker (and /resume latest) by the current workspace Cwd, checked BEFORE
the per-session read so a large global history no longer pays N full file
reads to build one workspace's list; also skip zero-event sessions via
metadata without any read. Sessions with no recorded Cwd (older runs) and
an unknown current workspace stay visible so nothing is hidden that can't
be confidently placed. Explicit /resume <id> still resolves any session.

* feat(tui): warn on model switch that drops staged images

Staged images were silently dropped at submit when the active model had
no vision support — the user only learned after sending. Chips already
render above the composer (renderAttachmentChips); add an immediate amber
warning line to the /model switch status (both same-provider and
cross-provider paths) when images are staged and the newly selected model
can't accept them, so the drop is surfaced at switch time.

* feat(tui): inline recovery affordance on setup errors

A first-run setup error rendered as a bare red 'error: <text>' with no
pointer to the fix. Append a faint, stage-tailored recovery line (edit
the endpoint/name/key above then Enter, or pick another model) that names
only keys the stage's footer confirms, so the guidance is always accurate.
Empty for stages whose footer already makes recovery obvious.

* fix: address CodeRabbit review (5 findings)

- loop.go: recompute the compaction calibration estimate at calibration
  time, not request-build time, so a reactive-compaction reissue that
  shrinks the request feeds the EMA the sample that actually completed.
- errhint: match bare HTTP status codes (401/403/429/529) only at digit
  boundaries, so an incidental number ('4290ms', 'id 14015') isn't
  mis-bucketed as auth/rate-limit.
- bash: drop the stray unmatched ']' in the output-truncation marker.
- tui /retry: apply the same exiting/compactInFlight guards commandPrompt
  has, so a retry can't race compactResultMsg's rewrite of session state.
- tui /export: write the transcript 0o600 (it may contain echoed
  secrets), not world/group-readable 0o644.

* test(tui): skip export file-mode assertion on Windows

os.WriteFile ignores Unix permission bits on Windows and Stat reports
0666, so the 0o600 assertion failed there. The production 0o600 is
honored on Linux/macOS (where it matters); guard the check with
runtime.GOOS so Windows smoke passes.

* fix(tui): align /resume latest filters + case-insensitive workspace match on Windows

Address CodeRabbit review on #21:
- latestResumableInWorkspace now applies the SAME filters as the picker
  (skip zero-event metadata and empty/failed runs), so /resume latest
  can't land on a session the picker intentionally hides.
- sessionMatchesWorkspace compares paths case-insensitively on Windows
  (its filesystem is), so the same workspace spelled with different casing
  no longer hides resumable sessions; other platforms stay case-sensitive.

* fix(tui): /retry resends the last prompt's image/PDF attachments

/retry relaunched only lastPrompt's text, but launchPrompt clears the
pending image/document queues once a turn is sent. A vision- or PDF-backed
prompt that failed would silently retry as text-only and answer a different
task. Snapshot the consumed attachments at launch and re-stage them on
/retry so launchPrompt rebuilds an identical request (document preamble,
images, and the submit-time vision re-check). startNewSession drops the
snapshot so a post-/new /retry cannot leak the previous session's
attachments. Also correct the OAuthResolver factory comment, which omitted
the Gemini provider it is now passed to.

* fix: address jatmn review (reconnect 5xx, /edit attachments, edit-cap, coalescer test)

- reconnect: exclude HTTP 5xx (500/502/503/504) before the transport
  substring match so "504 Gateway Timeout" no longer slips through on the
  generic "timeout" needle. A gateway timeout is non-idempotent — the
  completion POST may already have reached the model — so replaying the
  connect risked duplicate billable work. Reuses a digit-boundary status
  matcher exported from errhint (HasStatusCode). Test asserts 504/500 stay
  non-reconnectable.

- tui /edit: re-stage the remembered image/PDF snapshot alongside the
  recalled prompt text, mirroring the /retry fix — editing a vision- or
  document-backed prompt no longer silently resends a text-only version.

- compaction recent-edits: order edits by LAST edit, not first, and move a
  re-touched path to the newest position on merge, so the maxRecentEdits
  tail cap keeps the file the model most recently edited instead of dropping
  it. Adds mergeRecentEdits (edit-specific) and a capped-list regression.

- coalescer test: inject the frame timer via afterFunc so
  TestCoalescerBatchesDeltas proves buffering deterministically instead of
  racing the real 16ms wall-clock timer.

* fix(tools): bound bash output capture in memory, not just model-visible text

Previously stdout/stderr streamed into unbounded bytes.Buffers and were only
truncated to the 96 KiB budget AFTER the command finished, so a runaway command
(cat huge.log, yes) could grow Zero's memory until it stalled or OOMed before
budgetBashOutput ever ran. Replace the capture with boundedBuffer, an io.Writer
that keeps only the head+tail each stream will surface and discards the middle as
it arrives, while counting the true total for the truncation marker. Memory is now
bounded to ~head+2×tail per stream regardless of output size.

truncateHeadTail is refactored onto a total-aware core so the string path (and its
tests) are unchanged, and budgetBashCapture reports the true raw_bytes even though
only a bounded slice was held. Adds boundedBuffer/capture unit tests and an
end-to-end test that streams ~5× the budget and asserts bounded emission with an
accurate raw_bytes.
2026-07-05 13:15:14 +05:30

355 lines
12 KiB
Go

package tools
import (
"context"
"encoding/json"
"sort"
"strconv"
"strings"
)
// ToolSearchToolName is the canonical registry name of the deferred-tool loader.
// It is exported so the agent loop, CLI filter validation, and the partition can
// reference the loader without re-stating the literal (keeping all the gates that
// must treat it specially in agreement).
const ToolSearchToolName = "tool_search"
// toolSearchMaxKeywordMatches caps how many deferred tools a bare-keyword query
// loads in one call, keeping the returned schemas bounded.
const toolSearchMaxKeywordMatches = 10
// toolSearchTool lets the model pull a deferred tool's full schema on demand.
// It holds the live registry (like escalate_model holds the model registry) so
// it can resolve names against the currently registered deferred-eligible tools.
type toolSearchTool struct {
baseTool
registry *Registry
}
// NewToolSearchTool builds the tool_search tool over the given registry. The
// tool is informational/no-side-effect and is advertised even in auto mode so
// the model can always discover withheld tools.
func NewToolSearchTool(registry *Registry) Tool {
return toolSearchTool{
baseTool: baseTool{
name: ToolSearchToolName,
description: BuildToolSearchDescription(nil),
parameters: Schema{
Type: "object",
Properties: map[string]PropertySchema{
"query": {
Type: "string",
Description: "Either \"select:Name1,Name2\" for exact tool names, or space-separated keywords to match tool names and descriptions.",
},
},
Required: []string{"query"},
AdditionalProperties: false,
},
safety: Safety{
SideEffect: SideEffectNone,
Permission: PermissionAllow,
Reason: "Lists and loads already-registered tool schemas; performs no side effects.",
AdvertiseInAuto: true,
},
},
registry: registry,
}
}
// Run satisfies the Tool interface; actual dispatch goes through RunWithOptions.
func (tool toolSearchTool) Run(ctx context.Context, args map[string]any) Result {
return tool.RunWithOptions(ctx, args, RunOptions{})
}
func (tool toolSearchTool) RunWithOptions(_ context.Context, args map[string]any, options RunOptions) Result {
query, err := stringArg(args, "query", "", true)
if err != nil {
return errorResult("Error: Invalid arguments for tool_search: " + err.Error())
}
query = strings.TrimSpace(query)
// Honor the run's operator filters so an operator-hidden deferred tool is
// invisible to tool_search: it never resolves via select:, never ranks for a
// keyword query, and is omitted from the no-match listing.
deferred := tool.visibleDeferredTools(options.EnabledTools, options.DisabledTools)
var matches []Tool
if rest, ok := strings.CutPrefix(query, "select:"); ok {
matches = tool.resolveExact(rest, deferred)
} else {
matches = tool.rankByKeyword(query, deferred)
}
// Names the model asked for that it ALREADY has eagerly. tool_search only
// loads DEFERRED tools, so a request for an eager tool (e.g. Task or
// update_plan) otherwise falls through to a misleading "no tools matched" —
// a common confusion that leaves weaker models looping. Redirect the model to
// call those tools directly instead.
alreadyAvailable := tool.eagerToolsForQuery(query, options.EnabledTools, options.DisabledTools)
if len(matches) == 0 {
if len(alreadyAvailable) > 0 {
return okResult(alreadyAvailableMessage(alreadyAvailable))
}
return okResult(tool.noMatchMessage(query, deferred))
}
names := make([]string, 0, len(matches))
for _, match := range matches {
names = append(names, match.Name())
}
output := renderLoadedTools(matches)
if len(alreadyAvailable) > 0 {
output += "\n\n" + alreadyAvailableMessage(alreadyAvailable)
}
return Result{
Status: StatusOK,
Output: output,
Meta: map[string]string{"load_tools": strings.Join(names, ",")},
}
}
// visibleDeferredTools returns the registry's deferred-eligible tools that pass
// the operator allow/deny filters, sorted by name so keyword ranking and
// listings stay deterministic. A nil/empty filter pair admits every deferred
// tool (the pre-filter behavior). The allow/deny semantics mirror the agent's
// ToolAllowedByFilters: denied if in disabled; if enabled is non-empty, the tool
// must be listed in it.
func (tool toolSearchTool) visibleDeferredTools(enabled []string, disabled []string) []Tool {
var deferred []Tool
if tool.registry != nil {
for _, candidate := range tool.registry.All() {
if !IsDeferred(candidate) {
continue
}
if !toolAllowedByFilters(candidate.Name(), enabled, disabled) {
continue
}
deferred = append(deferred, candidate)
}
}
sort.Slice(deferred, func(left, right int) bool {
return deferred[left].Name() < deferred[right].Name()
})
return deferred
}
// visibleEagerToolNames returns the names of tools the model ALREADY has in its
// list this run: registered, NOT deferred, passing the operator filters, and not
// tool_search itself. These never need (and never resolve through) tool_search.
func (tool toolSearchTool) visibleEagerToolNames(enabled []string, disabled []string) map[string]bool {
names := map[string]bool{}
if tool.registry == nil {
return names
}
for _, candidate := range tool.registry.All() {
if IsDeferred(candidate) || candidate.Name() == ToolSearchToolName {
continue
}
if !toolAllowedByFilters(candidate.Name(), enabled, disabled) {
continue
}
names[candidate.Name()] = true
}
return names
}
// eagerToolsForQuery returns, sorted, the already-available (eager) tool names a
// tool_search query refers to: exact names for a "select:" query, or eager tools
// whose name matches a keyword otherwise. Used to steer the model back to calling
// a tool it already has instead of fruitlessly searching for it.
func (tool toolSearchTool) eagerToolsForQuery(query string, enabled []string, disabled []string) []string {
eager := tool.visibleEagerToolNames(enabled, disabled)
if len(eager) == 0 {
return nil
}
hit := map[string]bool{}
if rest, ok := strings.CutPrefix(query, "select:"); ok {
for _, raw := range strings.Split(rest, ",") {
if name := strings.TrimSpace(raw); name != "" && eager[name] {
hit[name] = true
}
}
} else {
keywords := strings.Fields(strings.ToLower(query))
for name := range eager {
lower := strings.ToLower(name)
for _, keyword := range keywords {
if strings.Contains(lower, keyword) {
hit[name] = true
break
}
}
}
}
out := make([]string, 0, len(hit))
for name := range hit {
out = append(out, name)
}
sort.Strings(out)
return out
}
// alreadyAvailableMessage tells the model the named tools are already in its list
// and must be called directly — tool_search is only for deferred tools.
func alreadyAvailableMessage(names []string) string {
subject, verb := names[0], "is"
if len(names) > 1 {
subject, verb = strings.Join(names, ", "), "are"
}
return subject + " " + verb + " already in your tool list — call " +
map[bool]string{true: "them", false: "it"}[len(names) > 1] +
" directly. tool_search only loads deferred tools, not tools you already have."
}
// toolAllowedByFilters mirrors agent.ToolAllowedByFilters (kept here to avoid an
// import cycle: the agent package imports tools, not the other way around). A
// name is denied when it appears in disabled; when enabled is non-empty, the
// name must appear in it to be allowed.
func toolAllowedByFilters(name string, enabled []string, disabled []string) bool {
if len(enabled) > 0 && !containsName(enabled, name) {
return false
}
return !containsName(disabled, name)
}
func containsName(names []string, name string) bool {
for _, candidate := range names {
if candidate == name {
return true
}
}
return false
}
// resolveExact maps a comma-separated name list (the part after "select:") to
// the matching deferred tools, preserving the model's order and skipping blanks
// and unknown names.
func (tool toolSearchTool) resolveExact(list string, deferred []Tool) []Tool {
byName := make(map[string]Tool, len(deferred))
for _, candidate := range deferred {
byName[candidate.Name()] = candidate
}
var matches []Tool
seen := make(map[string]bool)
for _, raw := range strings.Split(list, ",") {
name := strings.TrimSpace(raw)
if name == "" || seen[name] {
continue
}
if candidate, ok := byName[name]; ok {
seen[name] = true
matches = append(matches, candidate)
}
}
return matches
}
// rankByKeyword scores deferred tools by case-insensitive substring match on the
// name (weighted higher) then the description, and returns the top matches.
func (tool toolSearchTool) rankByKeyword(query string, deferred []Tool) []Tool {
keywords := strings.Fields(strings.ToLower(query))
if len(keywords) == 0 {
return nil
}
type scored struct {
tool Tool
score int
order int
}
var ranked []scored
for index, candidate := range deferred {
name := strings.ToLower(candidate.Name())
desc := strings.ToLower(candidate.Description())
nameSquashed := squashSeparators(name)
score := 0
for _, keyword := range keywords {
switch {
case strings.Contains(name, keyword):
score += 2
case strings.Contains(nameSquashed, squashSeparators(keyword)):
// Separator-insensitive fallback: "webfetch" matches "web_fetch".
// Ranked below an exact substring so precise queries still win.
score++
}
if strings.Contains(desc, keyword) {
score++
}
}
if score > 0 {
ranked = append(ranked, scored{tool: candidate, score: score, order: index})
}
}
sort.SliceStable(ranked, func(left, right int) bool {
if ranked[left].score != ranked[right].score {
return ranked[left].score > ranked[right].score
}
return ranked[left].order < ranked[right].order
})
if len(ranked) > toolSearchMaxKeywordMatches {
ranked = ranked[:toolSearchMaxKeywordMatches]
}
matches := make([]Tool, 0, len(ranked))
for _, entry := range ranked {
matches = append(matches, entry.tool)
}
return matches
}
// squashSeparators drops separators so "web_fetch", "web-fetch", and "web fetch"
// all normalize to "webfetch". It lets a query missing the separators still match
// a tool name, the common way a model mistypes a tool ("webfetch").
func squashSeparators(s string) string {
return strings.Map(func(r rune) rune {
switch r {
case '_', '-', ' ', '.':
return -1
}
return r
}, s)
}
// noMatchMessage reports that nothing loaded and names the available deferred
// tools so the model can retry with a valid select: query.
func (tool toolSearchTool) noMatchMessage(query string, deferred []Tool) string {
if len(deferred) == 0 {
return "No deferred tools are available to load."
}
names := make([]string, 0, len(deferred))
for _, candidate := range deferred {
names = append(names, candidate.Name())
}
return "No tools matched \"" + query + "\". Available tools: " + strings.Join(names, ", ") +
`. Retry with query "select:Name1,Name2" using exact names.`
}
// renderLoadedTools lists each loaded tool's name, full description, and full
// input schema (pretty-printed JSON) so the model has the complete spec inline.
func renderLoadedTools(matches []Tool) string {
var builder strings.Builder
builder.WriteString("Loaded ")
if len(matches) == 1 {
builder.WriteString("1 tool")
} else {
builder.WriteString(strconv.Itoa(len(matches)))
builder.WriteString(" tools")
}
builder.WriteString(". Full schemas follow; call them on the next turn.\n")
for _, match := range matches {
builder.WriteString("\n## ")
builder.WriteString(match.Name())
builder.WriteString("\n")
builder.WriteString(match.Description())
builder.WriteString("\n")
schemaJSON, err := json.MarshalIndent(match.Parameters(), "", " ")
if err != nil {
builder.WriteString("(schema unavailable)\n")
continue
}
builder.WriteString("input schema:\n")
builder.Write(schemaJSON)
builder.WriteString("\n")
}
return builder.String()
}