refactor(plugin): isolate HTTP hook dispatch

Co-authored-by: Aiden Cline <rekram1-node@users.noreply.github.com>

packages/ai/AGENTS.md

@@ -10,7 +10,7 @@
 
 ## Conventions
 
-Per-type constructors live on the type, not as top-level re-exports. Use `Message.system(...)`, `Message.user(...)`, `Message.assistant(...)`, `Message.tool(...)`, `LanguageModel.make(...)`, `ToolDefinition.make(...)`, `ToolCallPart.make(...)`, `ToolResultPart.make(...)`, `ToolChoice.make(...)`, `ToolChoice.named(...)`, `SystemPart.make(...)`, and `GenerationOptions.make(...)` directly. The top-level `LLM` namespace is reserved for request-shaped call APIs: `LLM.request`, `LLM.generate`, `LLM.stream`, `LLM.updateRequest`, and `LLM.generateObject`. Two ways to construct the same thing is one too many.
+Per-type constructors live on the type, not as top-level re-exports. Use `Message.system(...)`, `Message.user(...)`, `Message.assistant(...)`, `Message.tool(...)`, `LanguageModel.make(...)`, `ToolDefinition.make(...)`, `ToolCallPart.make(...)`, `ToolResultPart.make(...)`, `ToolChoice.make(...)`, `ToolChoice.named(...)`, `SystemPart.make(...)`, and `GenerationOptions.make(...)` directly. The top-level `LLM` namespace is reserved for request-shaped call APIs: `LLM.request`, `LLM.generate`, `LLM.stream`, and `LLM.generateObject`. Use `LLMRequest.update(...)` when deriving canonical request data; do not add a duplicate `LLM.updateRequest(...)` path. Two ways to construct the same thing is one too many.
 
 - Keep provider-defined string enums forward-compatible. Expose known values for autocomplete while accepting future values with `Known | (string & {})`; use `Schema.String` at runtime unless rejecting unknown values is required for correctness.
 

packages/ai/README.md

@@ -3,8 +3,9 @@
 Schema-first AI primitives for opencode. Provider quirks live in adapters, not in calling code.
 
 ```ts
-import { Effect } from "effect"
+import { Effect, Layer } from "effect"
 import { LLM, LLMClient } from "@opencode-ai/ai"
+import { RequestExecutor } from "@opencode-ai/ai/route"
 import { OpenAI } from "@opencode-ai/ai/providers"
 
 const model = OpenAI.configure({ apiKey: process.env.OPENAI_API_KEY }).responses("gpt-4o-mini")
@@ -20,6 +21,10 @@ const program = Effect.gen(function* () {
   const response = yield* LLMClient.generate(request)
   console.log(response.text)
 })
+
+const llmLayer = LLMClient.layer.pipe(Layer.provide(RequestExecutor.fetchLayer))
+
+await Effect.runPromise(program.pipe(Effect.provide(llmLayer)))
 ```
 
 Run `LLMClient.stream(request)` instead of `generate` when you want incremental `LLMEvent`s. The event stream is provider-neutral — same shape across OpenAI Chat, OpenAI Responses, Anthropic Messages, Gemini, Bedrock Converse, and any OpenAI-compatible deployment.
@@ -200,6 +205,32 @@ The hosted result is represented as a provider-executed tool call and tool resul
 - **`Image.generate({...})`** — generate images through a provider-neutral image request and response model.
 - **`ImageClient`** — Effect service and layer for image execution, parallel to `LLMClient`.
 
+## Testing
+
+Use the deterministic test client from `@opencode-ai/ai/testing` to script provider-neutral responses and inspect
+the requests sent by code under test:
+
+```ts
+import { Effect } from "effect"
+import { TestLLM } from "@opencode-ai/ai/testing"
+
+const testLLM = TestLLM.layer({
+  fallback: TestLLM.text("Hello from the test model", "text-1"),
+})
+
+// TestLLM.clientLayer provides LLMClient.Service and consumes TestLLM.Service.
+const programWithTestClient = Effect.gen(function* () {
+  const result = yield* program
+  const test = yield* TestLLM.Service
+  console.log(test.requests)
+  return result
+}).pipe(Effect.provide(TestLLM.clientLayer), Effect.provide(testLLM))
+```
+
+`TestLLM.push(...)` scripts one-shot responses, `TestLLM.always(...)` changes the fallback, and
+`TestLLM.wait(...)` lets concurrent tests wait until a request has arrived. Every received canonical request is
+available on the yielded `TestLLM.Service`.
+
 ## Caching
 
 Prompt caching is **on by default**. Every `LLMRequest` resolves to `cache: "auto"` unless the caller opts out with `cache: "none"`. Each protocol translates `CacheHint`s to its wire format (`cache_control` on Anthropic, `cachePoint` on Bedrock; OpenAI and Gemini do implicit caching server-side and don't need inline markers — auto is a no-op there).

packages/ai/src/image-client.ts

@@ -15,11 +15,11 @@ export class Service extends Context.Service<Service, Interface>()("@opencode/Im
 
 export const generate = <Options extends ImageOptions>(
   request: ImageRequestFor<Options>,
-): Effect.Effect<ImageResponse, AIError> =>
+): Effect.Effect<ImageResponse, AIError, Service> =>
   Effect.gen(function* () {
     const client = yield* Service
     return yield* client.generate(request)
-  }) as Effect.Effect<ImageResponse, AIError>
+  })
 
 export const layer: Layer.Layer<Service, never, RequestExecutor.Service> = Layer.effect(
   Service,

packages/ai/src/llm.ts

@@ -1,5 +1,5 @@
 import { Effect, JsonSchema, Schema } from "effect"
-import { LLMClient } from "./route/client"
+import { LLMClient, Service } from "./route/client"
 import {
   GenerationOptions,
   HttpOptions,
@@ -151,10 +151,10 @@ const runGenerateObject = Effect.fn("LLM.generateObject")(function* (
  */
 export function generateObject<const SelectedLanguageModel extends LanguageModel, S extends ToolSchema<any>>(
   options: GenerateObjectOptions<S, SelectedLanguageModel>,
-): Effect.Effect<GenerateObjectResponse<Schema.Schema.Type<S>>, AIError>
+): Effect.Effect<GenerateObjectResponse<Schema.Schema.Type<S>>, AIError, Service>
 export function generateObject<const SelectedLanguageModel extends LanguageModel>(
   options: GenerateObjectDynamicOptions<SelectedLanguageModel>,
-): Effect.Effect<GenerateObjectResponse<unknown>, AIError>
+): Effect.Effect<GenerateObjectResponse<unknown>, AIError, Service>
 export function generateObject(options: GenerateObjectOptions<ToolSchema<any>> | GenerateObjectDynamicOptions) {
   if ("schema" in options) {
     const { schema, ...rest } = options

packages/ai/src/route/client.ts

@@ -5,7 +5,7 @@ import { Endpoint, type EndpointPatch } from "./endpoint"
 import { RequestExecutor } from "./executor"
 import { Framing } from "./framing"
 import { HttpTransport } from "./transport"
-import type { HttpRequestTransform, Transport, TransportRuntime } from "./transport"
+import type { HttpMiddleware, Transport, TransportRuntime } from "./transport"
 import { WebSocketExecutor } from "./transport"
 import type { Protocol } from "./protocol"
 import { applyCachePolicy } from "../cache-policy"
@@ -155,7 +155,7 @@ export interface Interface {
 }
 
 export interface StreamOptions {
-  readonly transform?: HttpRequestTransform
+  readonly http?: HttpMiddleware
 }
 
 export interface StreamMethod {
@@ -307,7 +307,7 @@ function makeFromTransport<Body, Prepared, Frame, Event, State>(
           auth: routeInput.auth ?? Auth.none,
           encodeBody,
           headers: routeInput.headers,
-          transform: options?.transform,
+          middleware: options?.http,
         }),
       streamPrepared: (prepared: Prepared, request: LLMRequest, runtime: TransportRuntime) => {
         const route = `${request.model.provider}/${request.model.route.id}`
@@ -422,18 +422,18 @@ const generateWith = (stream: Interface["stream"]) =>
     )
   })
 
-export function stream(request: LLMRequest, options?: StreamOptions): Stream.Stream<LLMEvent, AIError> {
+export function stream(request: LLMRequest, options?: StreamOptions): Stream.Stream<LLMEvent, AIError, Service> {
   return Stream.unwrap(
     Effect.gen(function* () {
       return (yield* Service).stream(request, options)
     }),
-  ) as Stream.Stream<LLMEvent, AIError>
+  )
 }
 
-export function generate(request: LLMRequest, options?: StreamOptions): Effect.Effect<LLMResponse, AIError> {
+export function generate(request: LLMRequest, options?: StreamOptions): Effect.Effect<LLMResponse, AIError, Service> {
   return Effect.gen(function* () {
     return yield* (yield* Service).generate(request, options)
-  }) as Effect.Effect<LLMResponse, AIError>
+  })
 }
 
 export const streamRequest = (request: LLMRequest, options?: StreamOptions) =>

packages/ai/src/route/executor.ts

@@ -20,9 +20,18 @@ import { classifyProviderFailure } from "../provider-error"
 export interface Interface {
   readonly execute: (
     request: HttpClientRequest.HttpClientRequest,
+    middleware?: HttpMiddleware,
   ) => Effect.Effect<HttpClientResponse.HttpClientResponse, AIError>
 }
 
+export type HttpHandler = (
+  request: HttpClientRequest.HttpClientRequest,
+) => Effect.Effect<HttpClientResponse.HttpClientResponse, Error>
+export type HttpMiddleware = (
+  request: HttpClientRequest.HttpClientRequest,
+  handler: HttpHandler,
+) => Effect.Effect<HttpClientResponse.HttpClientResponse, Error>
+
 export class Service extends Context.Service<Service, Interface>()("@opencode/AI/RequestExecutor") {}
 
 const BODY_LIMIT = 16_384
@@ -261,7 +270,7 @@ const toHttpError = (redactedNames: ReadonlyArray<string | RegExp>) => (error: u
     return transportError({ message: error.message, kind: "Timeout" })
   }
   if (!HttpClientError.isHttpClientError(error)) {
-    return transportError({ message: "HTTP transport failed" })
+    return transportError({ message: error instanceof Error ? error.message : "HTTP transport failed" })
   }
   const request = "request" in error ? error.request : undefined
   if (error.reason._tag === "TransportError") {
@@ -282,12 +291,20 @@ export const layer: Layer.Layer<Service, never, HttpClient.HttpClient> = Layer.e
   Service,
   Effect.gen(function* () {
     const http = yield* HttpClient.HttpClient
-    const executeOnce = (request: HttpClientRequest.HttpClientRequest) =>
+    const executeOnce = (request: HttpClientRequest.HttpClientRequest, middleware?: HttpMiddleware) =>
       Effect.gen(function* () {
         const redactedNames = yield* Headers.CurrentRedactedNames
-        return yield* http
-          .execute(request)
-          .pipe(Effect.mapError(toHttpError(redactedNames)), Effect.flatMap(statusError(request, redactedNames)))
+        if (!middleware)
+          return yield* http
+            .execute(request)
+            .pipe(Effect.mapError(toHttpError(redactedNames)), Effect.flatMap(statusError(request, redactedNames)))
+
+        const response = yield* middleware(request, (input) =>
+          http
+            .execute(input)
+            .pipe(Effect.mapError((cause) => (cause instanceof Error ? cause : new Error(String(cause))))),
+        ).pipe(Effect.mapError(toHttpError(redactedNames)))
+        return yield* statusError(response.request, redactedNames)(response)
       })
     return Service.of({
       execute: executeOnce,

packages/ai/src/route/index.ts

@@ -23,4 +23,4 @@ export type { ApiKeyMode, AuthOverride, ProviderAuthOption } from "./auth-option
 export type { Definition as EndpointFn, EndpointInput } from "./endpoint"
 export type { Definition as FramingDef } from "./framing"
 export type { Protocol as ProtocolDef } from "./protocol"
-export type { HttpRequest, HttpRequestTransform, Transport as TransportDef, TransportRuntime } from "./transport"
+export type { HttpHandler, HttpMiddleware, Transport as TransportDef, TransportRuntime } from "./transport"

packages/ai/src/route/transport/http.ts

@@ -3,7 +3,7 @@ import { Headers, HttpClientRequest } from "effect/unstable/http"
 import { Auth } from "../auth"
 import { render as renderEndpoint } from "../endpoint"
 import { Framing } from "../framing"
-import type { Transport, TransportPrepareInput } from "./index"
+import type { HttpMiddleware, Transport, TransportPrepareInput } from "./index"
 import * as ProviderShared from "../../protocols/shared"
 import { mergeJsonRecords, type LLMRequest } from "../../schema"
 
@@ -19,6 +19,7 @@ export interface JsonRequestParts<Body = unknown> {
 export interface HttpPrepared<Frame> {
   readonly request: HttpClientRequest.HttpClientRequest
   readonly framing: Framing.Definition<Frame>
+  readonly middleware?: HttpMiddleware
 }
 
 const applyQuery = (url: string, query: Record<string, string> | undefined) => {
@@ -74,21 +75,21 @@ export const httpJson = <Body, Frame>(input: HttpJsonInput<Body, Frame>): HttpJs
   prepare: (prepareInput) =>
     Effect.gen(function* () {
       const parts = yield* jsonRequestParts({ ...prepareInput })
-      const request = { url: parts.url, method: "POST", headers: { ...parts.headers }, body: parts.bodyText }
-      yield* (prepareInput.transform?.(request) ?? Effect.void)
+      const request = ProviderShared.jsonPost({
+        url: parts.url,
+        body: parts.bodyText,
+        headers: parts.headers,
+      })
       return {
-        request: ProviderShared.jsonPost({
-          url: request.url,
-          body: request.body ?? "",
-          headers: Headers.fromInput(request.headers),
-        }),
+        request,
         framing: input.framing,
+        middleware: prepareInput.middleware,
       }
     }),
   frames: (prepared, request, runtime) =>
     Stream.unwrap(
       runtime.http
-        .execute(prepared.request)
+        .execute(prepared.request, prepared.middleware)
         .pipe(
           Effect.map((response) =>
             prepared.framing.frame(

packages/ai/src/route/transport/index.ts

@@ -1,7 +1,7 @@
 import type { Effect, Stream } from "effect"
 import { Endpoint } from "../endpoint"
 import { Auth } from "../auth"
-import type { Interface as RequestExecutorInterface } from "../executor"
+import type { HttpMiddleware, Interface as RequestExecutorInterface } from "../executor"
 import type { Interface as WebSocketExecutorInterface } from "./websocket"
 import type { AIError, LLMRequest } from "../../schema"
 
@@ -10,15 +10,6 @@ export interface TransportRuntime {
   readonly webSocket?: WebSocketExecutorInterface
 }
 
-export interface HttpRequest {
-  url: string
-  readonly method: string
-  headers: Record<string, string>
-  body: string | undefined
-}
-
-export type HttpRequestTransform = (request: HttpRequest) => Effect.Effect<void>
-
 export interface Transport<Body, Prepared, Frame> {
   readonly id: string
   readonly prepare: (input: TransportPrepareInput<Body>) => Effect.Effect<Prepared, AIError>
@@ -32,8 +23,9 @@ export interface TransportPrepareInput<Body> {
   readonly auth: Auth.Definition
   readonly encodeBody: (body: Body) => string
   readonly headers?: (input: { readonly request: LLMRequest }) => Record<string, string>
-  readonly transform?: HttpRequestTransform
+  readonly middleware?: HttpMiddleware
 }
 
 export * as HttpTransport from "./http"
+export type { HttpHandler, HttpMiddleware } from "../executor"
 export { WebSocketExecutor, WebSocketTransport } from "./websocket"

packages/ai/test/compile.test.ts

@@ -1,6 +1,6 @@
 import { describe, expect, test } from "bun:test"
-import { Effect, Schema } from "effect"
-import { HttpClientRequest } from "effect/unstable/http"
+import { Effect, Ref, Schema } from "effect"
+import { HttpClientRequest, HttpClientResponse } from "effect/unstable/http"
 import { LLM, mergeProviderOptions } from "../src"
 import { AnthropicMessages, OpenAIChat } from "../src/protocols"
 import { Auth, LLMClient } from "../src/route"
@@ -146,12 +146,16 @@ describe("request option precedence", () => {
         prompt: "Say hello.",
       }),
       {
-        transform: (request) =>
-          Effect.sync(() => {
-            expect(request.headers.authorization).toBe("Bearer fresh-key")
-            request.url = "https://proxy.test/v1/chat/completions"
-            request.headers["x-plugin"] = "transformed"
-            request.body = JSON.stringify({ transformed: true })
+        http: (request, handler) =>
+          Effect.gen(function* () {
+            return yield* handler(
+              request.pipe(
+                HttpClientRequest.setUrl("https://proxy.test/v1/chat/completions"),
+                HttpClientRequest.setMethod("PUT"),
+                HttpClientRequest.setHeader("x-plugin", "transformed"),
+                HttpClientRequest.bodyText(JSON.stringify({ transformed: true }), "application/custom+json"),
+              ),
+            )
           }),
       },
     ).pipe(
@@ -160,7 +164,9 @@ describe("request option precedence", () => {
           Effect.gen(function* () {
             const web = yield* HttpClientRequest.toWeb(input.request).pipe(Effect.orDie)
             expect(web.url).toBe("https://proxy.test/v1/chat/completions")
+            expect(web.method).toBe("PUT")
             expect(web.headers.get("x-plugin")).toBe("transformed")
+            expect(web.headers.get("content-type")).toBe("application/custom+json")
             expect(decodeJson(input.text)).toEqual({ transformed: true })
             return input.respond(sseEvents(deltaChunk({}, "stop")), {
               headers: { "content-type": "text/event-stream" },
@@ -171,6 +177,82 @@ describe("request option precedence", () => {
     ),
   )
 
+  it.effect("transforms the HTTP response before protocol decoding", () =>
+    Effect.gen(function* () {
+      const response = yield* LLMClient.generate(
+        LLM.request({
+          model: OpenAIChat.route
+            .with({ endpoint: { baseURL: "https://api.openai.test/v1/" }, auth: Auth.bearer("test") })
+            .model({ id: "gpt-4o-mini" }),
+          prompt: "Say hello.",
+        }),
+        {
+          http: (request, handler) =>
+            Effect.gen(function* () {
+              const response = yield* handler(request)
+              return HttpClientResponse.fromWeb(
+                response.request,
+                new Response((yield* response.text).replace("network", "hooked"), {
+                  status: response.status,
+                  headers: response.headers,
+                }),
+              )
+            }),
+        },
+      ).pipe(
+        Effect.provide(
+          dynamicResponse((input) =>
+            Effect.succeed(
+              input.respond(sseEvents(deltaChunk({ content: "network" }, "stop")), {
+                headers: { "content-type": "text/event-stream" },
+              }),
+            ),
+          ),
+        ),
+      )
+
+      expect(response.text).toBe("hooked")
+    }),
+  )
+
+  it.effect("can inspect an error response and retry the native request", () =>
+    Effect.gen(function* () {
+      const attempts = yield* Ref.make(0)
+      const response = yield* LLMClient.generate(
+        LLM.request({
+          model: OpenAIChat.route
+            .with({ endpoint: { baseURL: "https://api.openai.test/v1/" }, auth: Auth.bearer("stale") })
+            .model({ id: "gpt-4o-mini" }),
+          prompt: "Say hello.",
+        }),
+        {
+          http: (request, handler) =>
+            Effect.gen(function* () {
+              const response = yield* handler(request)
+              expect(response.status).toBe(401)
+              return yield* handler(HttpClientRequest.setHeader(request, "authorization", "Bearer refreshed"))
+            }),
+        },
+      ).pipe(
+        Effect.provide(
+          dynamicResponse((input) =>
+            Effect.gen(function* () {
+              yield* Ref.update(attempts, (value) => value + 1)
+              if (input.request.headers.authorization !== "Bearer refreshed")
+                return input.respond("unauthorized", { status: 401 })
+              return input.respond(sseEvents(deltaChunk({ content: "retried" }, "stop")), {
+                headers: { "content-type": "text/event-stream" },
+              })
+            }),
+          ),
+        ),
+      )
+
+      expect(response.text).toBe("retried")
+      expect(yield* Ref.get(attempts)).toBe(2)
+    }),
+  )
+
   it.effect("applies raw body overlays after protocol lowering", () =>
     LLMClient.generate(
       LLM.request({

packages/ai/test/executor.test.ts

@@ -67,6 +67,18 @@ const expectAIError = (error: unknown) => {
 const errorHttp = (error: AIError) => ("http" in error.reason ? error.reason.http : undefined)
 
 describe("RequestExecutor", () => {
+  it.effect("preserves middleware error messages", () =>
+    Effect.gen(function* () {
+      const executor = yield* RequestExecutor.Service
+      const error = yield* executor
+        .execute(request, () => Effect.fail(new Error("plugin rejected request")))
+        .pipe(Effect.flip)
+
+      expectAIError(error)
+      expect(error.reason.message).toBe("plugin rejected request")
+    }).pipe(Effect.provide(responsesLayer([]))),
+  )
+
   it.effect("classifies context overflow responses", () =>
     Effect.gen(function* () {
       const executor = yield* RequestExecutor.Service

packages/ai/test/image.types.ts

@@ -1,5 +1,7 @@
+import { Effect } from "effect"
 import {
   Image,
+  ImageClient,
   ImageInput,
   ImageModel,
   type ImageModelOptions,
@@ -7,8 +9,13 @@ import {
   type ImageRequestFor,
   type ImageRoute,
 } from "../src"
+import type { Service } from "../src/image-client"
 import { Google, OpenAI, XAI, ZAI } from "../src/providers"
 
+type Requirements<T> = T extends Effect.Effect<infer _A, infer _E, infer R> ? R : never
+type Equal<A, B> = [A, B] extends [B, A] ? true : false
+type Assert<T extends true> = T
+
 type GoogleLikeOptions = {
   readonly aspectRatio?: "1:1" | "16:9"
   readonly imageSize?: "1K" | "2K"
@@ -146,6 +153,9 @@ const request = Image.request({
 })
 const typedRequest: ImageRequestFor<GoogleLikeOptions> = request
 void typedRequest
+const generated = ImageClient.generate(request)
+type GenerateRequirements = Assert<Equal<Requirements<typeof generated>, Service>>
+void (true satisfies GenerateRequirements)
 
 // @ts-expect-error Image requests no longer expose a common count option.
 Image.generate({ model: openai, prompt: "A lighthouse", count: 2 })

packages/ai/test/llm-option-types.types.ts

@@ -1,5 +1,11 @@
-import { Schema } from "effect"
-import { LLM, type LanguageModel, type LanguageModelProviderOptions, type ProviderOptions } from "../src"
+import { Effect, Schema, Stream } from "effect"
+import {
+  LLM,
+  type LLMClientService,
+  type LanguageModel,
+  type LanguageModelProviderOptions,
+  type ProviderOptions,
+} from "../src"
 import { OpenAIChat } from "../src/protocols"
 
 interface ExampleOptions {
@@ -15,9 +21,19 @@ const model = OpenAIChat.route
   .with({ endpoint: { baseURL: "https://example.com/v1" } })
   .model<ExampleProviderOptions>({ id: "example" })
 
+type Requirements<T> = T extends Effect.Effect<infer _A, infer _E, infer R> ? R : never
+type StreamRequirements<T> = T extends Stream.Stream<infer _A, infer _E, infer R> ? R : never
+type Equal<A, B> = [A, B] extends [B, A] ? true : false
+type Assert<T extends true> = T
+
 LLM.request({ model, prompt: "Hello", providerOptions: { example: { mode: "fast" } } })
 LLM.request({ model, prompt: "Hello", providerOptions: { future: { option: true } } })
 
+const generated = LLM.generate(LLM.request({ model, prompt: "Hello" }))
+type GenerateRequirements = Assert<Equal<Requirements<typeof generated>, LLMClientService>>
+const streamed = LLM.stream(LLM.request({ model, prompt: "Hello" }))
+type StreamClientRequirements = Assert<Equal<StreamRequirements<typeof streamed>, LLMClientService>>
+
 LLM.request({
   model,
   prompt: "Hello",
@@ -25,12 +41,20 @@ LLM.request({
   providerOptions: { example: { mode: "slow" } },
 })
 
-LLM.generateObject({
+const generatedObject = LLM.generateObject({
   model,
   prompt: "Hello",
   schema: Schema.Struct({ answer: Schema.String }),
   providerOptions: { example: { mode: "thorough" } },
 })
+type GenerateObjectRequirements = Assert<Equal<Requirements<typeof generatedObject>, LLMClientService>>
+
+const generatedDynamicObject = LLM.generateObject({
+  model,
+  prompt: "Hello",
+  jsonSchema: { type: "object" },
+})
+type GenerateDynamicObjectRequirements = Assert<Equal<Requirements<typeof generatedDynamicObject>, LLMClientService>>
 
 LLM.generateObject({
   model,
@@ -44,4 +68,8 @@ declare const generic: LanguageModel
 LLM.request({ model: generic, prompt: "Hello", providerOptions: { arbitrary: { option: true } } })
 
 const options: LanguageModelProviderOptions<typeof model> = { example: { mode: "fast" } }
-void options
+void (options satisfies LanguageModelProviderOptions<typeof model>)
+void (true satisfies GenerateRequirements)
+void (true satisfies StreamClientRequirements)
+void (true satisfies GenerateObjectRequirements)
+void (true satisfies GenerateDynamicObjectRequirements)

packages/core/src/agent.ts

@@ -61,7 +61,7 @@ const layer = Layer.effect(
           draft.default = id
         },
         update: (id, fn) => {
-          const current = draft.agents.get(id) ?? (Info.empty(id) as Types.DeepMutable<Info>)
+          const current = draft.agents.get(id) ?? (Info.default(id) as Types.DeepMutable<Info>)
           if (!draft.agents.has(id)) draft.agents.set(id, current)
           fn(current)
           current.id = id

packages/core/src/config/plugin/agent.ts

@@ -58,7 +58,7 @@ export const Plugin = define({
           const files = yield* discover(fs, entry.path)
           return yield* Effect.forEach(files, (file) =>
             fs.readFileStringSafe(file.filepath).pipe(
-              Effect.map((content) => content && decode(file, content)),
+              Effect.map((content) => (content ? decode(file, content) : undefined)),
               Effect.catch(() => Effect.succeed(undefined)),
             ),
           ).pipe(

packages/core/src/plugin/hooks.ts

@@ -1,13 +1,22 @@
 export * as PluginHooks from "./hooks"
 
 import type { AISDKHooks } from "@opencode-ai/plugin/effect/aisdk"
-import type { SessionHooks } from "@opencode-ai/plugin/effect/session"
+import type { SessionContext, SessionHttpContext } from "@opencode-ai/plugin/effect/session"
 import type { ShellHooks } from "@opencode-ai/plugin/effect/shell"
 import type { ToolHooks } from "@opencode-ai/plugin/effect/tool"
 import { Context, Effect, Layer, Scope } from "effect"
 import { makeLocationNode } from "@opencode-ai/util/effect/app-node"
 import { State } from "../state"
 
+interface SessionHttpEvent extends SessionHttpContext {
+  request: (input: Request) => Effect.Effect<Response, Error>
+}
+
+interface SessionHooks {
+  readonly context: SessionContext
+  readonly http: SessionHttpEvent
+}
+
 export interface Domains {
   readonly aisdk: AISDKHooks
   readonly session: SessionHooks

packages/core/src/plugin/host.ts

@@ -2,6 +2,7 @@ export * as PluginHost from "./host"
 
 import { Plugin } from "@opencode-ai/plugin/effect"
 import type { IntegrationMethodRegistration } from "@opencode-ai/plugin/effect/integration"
+import type { SessionHookRegistration } from "@opencode-ai/plugin/effect/session"
 import type { CredentialOAuth } from "@opencode-ai/sdk/v2/types"
 import { EventManifest } from "@opencode-ai/schema/event-manifest"
 import { App } from "../app"
@@ -337,7 +338,16 @@ export const make = Effect.fn("PluginHost.make")(function* (plugin: import("../p
         }),
     },
     session: {
-      hook: (name, callback) => hooks.register("session", name, callback),
+      hook: (...registration: SessionHookRegistration) => {
+        if (registration[0] !== "http") return hooks.register("session", ...registration)
+        const middleware = registration[1]
+        return hooks.register("session", "http", (event) =>
+          Effect.sync(() => {
+            const next = event.request
+            event.request = (request) => middleware(event, request, next)
+          }),
+        )
+      },
       create: (input) =>
         runtime.session.create({
           id: input?.id,

packages/core/src/plugin/promise.ts

@@ -2,6 +2,7 @@ export * as PluginPromise from "./promise"
 
 import { define } from "@opencode-ai/plugin/effect/plugin"
 import type { Context, Plugin } from "@opencode-ai/plugin/promise/plugin"
+import type { SessionHookRegistration } from "@opencode-ai/plugin/promise/session"
 import type { Info } from "@opencode-ai/plugin/promise/tool"
 import { Agent } from "@opencode-ai/schema/agent"
 import { Integration } from "@opencode-ai/schema/integration"
@@ -194,7 +195,9 @@ export function fromPromise(plugin: Plugin) {
                               ),
                             ),
                           refresh:
-                            refresh === undefined ? undefined : (credential) => Effect.promise(() => refresh(credential)),
+                            refresh === undefined
+                              ? undefined
+                              : (credential) => Effect.promise(() => refresh(credential)),
                         })
                       },
                       remove: draft.method.remove,
@@ -263,8 +266,28 @@ export function fromPromise(plugin: Plugin) {
               ),
           },
           session: {
-            hook: (name, callback) =>
-              register(host.session.hook(name, (event) => Effect.promise(() => Promise.resolve(callback(event))))),
+            hook: (...registration: SessionHookRegistration) => {
+              if (registration[0] !== "http")
+                return register(
+                  host.session.hook(registration[0], (event) =>
+                    Effect.promise(() => Promise.resolve(registration[1](event))),
+                  ),
+                )
+              const middleware = registration[1]
+              return register(
+                host.session.hook("http", (event, input, next) =>
+                  Effect.tryPromise({
+                    try: (signal) =>
+                      Promise.resolve(
+                        middleware(event, new Request(input, { signal }), (request) =>
+                          Effect.runPromiseWith(context)(next(new Request(request, { signal })), { signal }),
+                        ),
+                      ),
+                    catch: (cause) => (cause instanceof Error ? cause : new Error(String(cause))),
+                  }),
+                ),
+              )
+            },
             create: (input) =>
               run(
                 host.session.create(

packages/core/src/plugin/provider/openai.ts

@@ -225,17 +225,14 @@ export const OpenAIPlugin = define({
         })
       }
     })
-    yield* ctx.session.hook("request", (evt) =>
-      Effect.sync(() => {
-        if (!chatgpt || evt.model.providerID !== Provider.ID.openai) return
-        const url = new URL(evt.url)
-        if (url.origin === "https://api.openai.com") {
-          evt.url = `${codexBaseURL}${url.pathname.replace(/^\/v1/, "")}${url.search}`
-        }
-        evt.headers.originator = "opencode"
-        evt.headers["session-id"] = evt.sessionID
-      }),
-    )
+    yield* ctx.session.hook("http", (evt, request, next) => {
+      if (!chatgpt || evt.model.providerID !== Provider.ID.openai) return next(request)
+      const url = new URL(request.url)
+      request.headers.set("originator", "opencode")
+      request.headers.set("session-id", evt.sessionID)
+      if (url.origin !== "https://api.openai.com") return next(request)
+      return next(new Request(`${codexBaseURL}${url.pathname.replace(/^\/v1/, "")}${url.search}`, request))
+    })
 
     const refresh = () => loading.withPermit(load().pipe(Effect.andThen(ctx.catalog.reload())))
     yield* bus.subscribe(Integration.Event.ConnectionUpdated).pipe(

packages/core/src/session/model-request.ts

@@ -4,7 +4,8 @@ import { LLM, Message, SystemPart, type LLMRequest } from "@opencode-ai/ai"
 import type { StreamOptions } from "@opencode-ai/ai/route"
 import type { Content } from "@opencode-ai/schema/tool"
 import { SessionError } from "@opencode-ai/schema/session-error"
-import { Cause, Config, Context, Effect, Layer, Result } from "effect"
+import { Cause, Config, Context, Effect, Layer, Result, Stream } from "effect"
+import { HttpClientRequest, HttpClientResponse } from "effect/unstable/http"
 import { makeLocationNode } from "@opencode-ai/util/effect/app-node"
 import { App } from "../app"
 import { Model } from "../model"
@@ -48,9 +49,7 @@ interface Prepared {
    * One request-scoped execution operation. Unknown, hook-removed, and
    * step-limit-violating calls fail individually through the same seam.
    */
-  readonly executeTool: (
-    input: Parameters<Tool.Snapshot["execute"]>[0],
-  ) => Effect.Effect<Tool.Result, ExecuteError>
+  readonly executeTool: (input: Parameters<Tool.Snapshot["execute"]>[0]) => Effect.Effect<Tool.Result, ExecuteError>
   /** True when this request is the final Step; violating calls are rejected and no continuation follows. */
   readonly stepLimitReached: boolean
 }
@@ -137,8 +136,7 @@ export const boundImages = (messages: LLMRequest["messages"]) => {
           result: {
             ...part.result,
             value: part.result.value.map((item: Content) => {
-              if (item.type !== "file" || !isImage(item.mime) || imageBytes - removed <= IMAGE_BYTES_TARGET)
-                return item
+              if (item.type !== "file" || !isImage(item.mime) || imageBytes - removed <= IMAGE_BYTES_TARGET) return item
               removed += Buffer.byteLength(item.uri)
               return { type: "text" as const, text: IMAGE_REMOVED }
             }),
@@ -204,9 +202,7 @@ export const layer = Layer.effect(
       })
       const hookedTools = Object.entries(contextEvent.tools).flatMap(([name, tool]) => {
         const registered = toolsByName.get(name)
-        return registered
-          ? [{ ...registered, description: tool.description, inputSchema: tool.input }]
-          : []
+        return registered ? [{ ...registered, description: tool.description, inputSchema: tool.input }] : []
       })
       const request = LLM.request({
         model,
@@ -220,24 +216,37 @@ export const layer = Layer.effect(
         toolChoice: stepLimitReached ? "none" : undefined,
       })
       const options: StreamOptions = {
-        transform: (request) =>
-          hooks
-            .trigger("session", "request", {
+        http: (request, handler) =>
+          Effect.gen(function* () {
+            let sent = request
+            const origins = new WeakMap<Response, HttpClientRequest.HttpClientRequest>()
+            const web = yield* HttpClientRequest.toWeb(request)
+            const event = yield* hooks.trigger("session", "http", {
               sessionID: session.id,
               agent: agent.id,
               model: resolved.ref,
-              ...request,
-            })
-            .pipe(
-              Effect.tap((event) =>
-                Effect.sync(() => {
-                  request.url = event.url
-                  request.headers = event.headers
-                  request.body = event.body
+              request: (input) =>
+                Effect.gen(function* () {
+                  sent = HttpClientRequest.fromWeb(input)
+                  if (input.body)
+                    sent = HttpClientRequest.bodyUint8Array(
+                      sent,
+                      new Uint8Array(yield* Effect.promise(() => input.clone().arrayBuffer())),
+                      input.headers.get("content-type") ?? undefined,
+                    )
+                  const response = yield* handler(sent)
+                  const body = [204, 205, 304].includes(response.status)
+                    ? null
+                    : yield* Stream.toReadableStreamEffect(response.stream)
+                  const output = new Response(body, { status: response.status, headers: response.headers })
+                  origins.set(output, sent)
+                  return output
                 }),
-              ),
-              Effect.asVoid,
-            ),
+            })
+            const response = yield* event.request(web)
+            const origin = origins.get(response) ?? sent
+            return HttpClientResponse.fromWeb(origin, response)
+          }).pipe(Effect.mapError((cause) => (cause instanceof Error ? cause : new Error(String(cause))))),
       }
       if (promptCacheSnapshots) {
         const current = PromptCacheDiagnostics.snapshot(request)
@@ -257,8 +266,7 @@ export const layer = Layer.effect(
         )
       }
       const executeTool: Prepared["executeTool"] = (executeInput) => {
-        if (stepLimitReached)
-          return new Tool.Error({ message: "Tools are disabled after the maximum agent steps" })
+        if (stepLimitReached) return new Tool.Error({ message: "Tools are disabled after the maximum agent steps" })
         if (toolsByName.has(executeInput.call.name) && !Object.hasOwn(contextEvent.tools, executeInput.call.name))
           return new Tool.Error({ message: `Tool is not available for this request: ${executeInput.call.name}` })
         return tools

packages/core/test/agent.test.ts

@@ -120,7 +120,7 @@ describe("Agent", () => {
       const id = Agent.ID.make("custom")
 
       yield* agent.transform((editor) => editor.update(id, () => {}))
-      expect(yield* agent.get(id)).toEqual(Agent.Info.empty(id))
+      expect(yield* agent.get(id)).toEqual(Agent.Info.default(id))
 
       yield* agent.transform((editor) => editor.remove(id))
       expect(yield* agent.get(id)).toBeUndefined()

packages/core/test/config/agent.test.ts

@@ -266,6 +266,7 @@ permissions:
 Use native v2 fields.`,
             )
             await fs.writeFile(path.join(tmp.path, "agents", "disabled.md"), "---\ndisabled: true\n---\nDisabled")
+            await fs.writeFile(path.join(tmp.path, "agents", "empty.md"), "")
             await fs.writeFile(path.join(tmp.path, "modes", "plan.md"), "Make a plan.")
           })
           const agents = yield* Agent.Service
@@ -295,6 +296,7 @@ Use native v2 fields.`,
             permissions: [...defaultPermissions, { action: "edit", resource: "*", effect: "deny" }],
           })
           expect(yield* agents.get(Agent.ID.make("disabled"))).toBeUndefined()
+          expect(yield* agents.get(Agent.ID.make("empty"))).toBeUndefined()
           expect(yield* agents.get(Agent.ID.make("plan"))).toMatchObject({ system: "Make a plan.", mode: "primary" })
         }),
       ),

packages/core/test/mcp-instructions.test.ts

@@ -12,7 +12,7 @@ import { readInitial, readUpdate } from "./lib/instructions"
 const build = Agent.ID.make("build")
 
 const selection = (permissions: Permission.Ruleset = []) => {
-  const info = Agent.Info.make({ ...Agent.Info.empty(build), permissions })
+  const info = Agent.Info.make({ ...Agent.Info.default(build), permissions })
   return { id: info.id, info }
 }
 

packages/core/test/plugin/promise.test.ts

@@ -1,6 +1,6 @@
 import { describe, expect } from "bun:test"
 import { Message, SystemPart } from "@opencode-ai/ai"
-import { DateTime, Effect, Schema } from "effect"
+import { DateTime, Deferred, Effect, Fiber, Schema } from "effect"
 import { Agent } from "@opencode-ai/core/agent"
 import { Catalog } from "@opencode-ai/core/catalog"
 import { Model } from "@opencode-ai/core/model"
@@ -148,7 +148,9 @@ describe("fromPromise", () => {
             expect((await ctx.agent.get({ agentID: Agent.ID.make("reviewer") })).data).toMatchObject({
               description: "Reviews code",
             })
-            await expect(ctx.agent.get({ agentID: Agent.ID.make("missing") })).rejects.toThrow("Agent not found: missing")
+            await expect(ctx.agent.get({ agentID: Agent.ID.make("missing") })).rejects.toThrow(
+              "Agent not found: missing",
+            )
             const models = (await ctx.catalog.model.list()).data
             expect(models.find((model) => model.providerID === "test" && model.id === "alias")).toMatchObject({
               modelID: "gpt-5",
@@ -221,6 +223,76 @@ describe("fromPromise", () => {
     }),
   )
 
+  it.effect("adapts promise session HTTP hooks", () =>
+    Effect.gen(function* () {
+      const plugin = yield* Plugin.Service
+      const hooks = yield* PluginHooks.Service
+      const host = yield* PluginHost.make(plugin)
+      yield* PluginPromise.fromPromise(
+        define({
+          id: "promise-session-http",
+          setup: async (ctx) => {
+            await ctx.session.hook("http", async (_event, request, next) => {
+              request.headers.set("x-hook", "promise")
+              const response = await next(request)
+              return new Response(`${await response.text()}-response`)
+            })
+            await ctx.session.hook("http", async (_event, request, next) => {
+              const response = await next(request)
+              return new Response(`${await response.text()}-outer`)
+            })
+          },
+        }),
+      ).effect(host)
+      const event: PluginHooks.Domains["session"]["http"] = {
+        sessionID: Session.ID.make("ses_promise_session_http"),
+        agent: Agent.ID.make("build"),
+        model: Model.Ref.make({ providerID: Provider.ID.make("test"), id: Model.ID.make("model") }),
+        request: (input) => Effect.succeed(new Response(input.headers.get("x-hook") ?? "missing")),
+      }
+
+      yield* hooks.trigger("session", "http", event)
+      const response = yield* event.request(new Request("https://provider.test"))
+
+      expect(yield* Effect.promise(() => response.text())).toBe("promise-response-outer")
+    }),
+  )
+
+  it.effect("interrupts the Effect request through a promise session HTTP hook", () =>
+    Effect.gen(function* () {
+      const plugin = yield* Plugin.Service
+      const hooks = yield* PluginHooks.Service
+      const host = yield* PluginHost.make(plugin)
+      yield* PluginPromise.fromPromise(
+        define({
+          id: "promise-session-http-interrupt",
+          setup: async (ctx) => {
+            await ctx.session.hook("http", (_event, request, next) => next(request))
+          },
+        }),
+      ).effect(host)
+      const started = yield* Deferred.make<void>()
+      const interrupted = yield* Deferred.make<void>()
+      const event: PluginHooks.Domains["session"]["http"] = {
+        sessionID: Session.ID.make("ses_promise_session_http_interrupt"),
+        agent: Agent.ID.make("build"),
+        model: Model.Ref.make({ providerID: Provider.ID.make("test"), id: Model.ID.make("model") }),
+        request: () =>
+          Deferred.succeed(started, undefined).pipe(
+            Effect.andThen(Effect.never),
+            Effect.onInterrupt(() => Deferred.succeed(interrupted, undefined)),
+          ),
+      }
+
+      yield* hooks.trigger("session", "http", event)
+      const fiber = yield* event.request(new Request("https://provider.test")).pipe(Effect.forkChild)
+      yield* Deferred.await(started)
+      yield* Fiber.interrupt(fiber)
+
+      expect(yield* Deferred.isDone(interrupted)).toBeTrue()
+    }),
+  )
+
   it.effect("disposes a hook registration on request", () =>
     Effect.gen(function* () {
       const agents = yield* Agent.Service
@@ -315,19 +387,17 @@ describe("fromPromise", () => {
         id: "promise-tool",
         setup: async (ctx) => {
           await ctx.tool.transform((tools) => {
-            tools.add(
-              {
-                name: "hello",
-                options: { codemode: false },
-                description: "Hello",
-                input: Schema.Struct({ name: Schema.String }),
-                output: Schema.String,
-                execute: async ({ name }, context) => {
-                  await context.progress({ phase: "greeting" })
-                  return { output: `Hello, ${name}!` }
-                },
+            tools.add({
+              name: "hello",
+              options: { codemode: false },
+              description: "Hello",
+              input: Schema.Struct({ name: Schema.String }),
+              output: Schema.String,
+              execute: async ({ name }, context) => {
+                await context.progress({ phase: "greeting" })
+                return { output: `Hello, ${name}!` }
               },
-            )
+            })
           })
         },
       })

packages/core/test/plugin/provider-openai.test.ts

@@ -29,6 +29,21 @@ function required<T>(value: T | undefined): T {
   return value
 }
 
+const http = Effect.fn(function* (providerID: Provider.ID, url: string) {
+  const event = yield* (yield* PluginHooks.Service).trigger("session", "http", {
+    sessionID: Session.ID.make("ses_test"),
+    agent: Agent.ID.make("build"),
+    model: Model.Ref.make({ providerID, id: Model.ID.make("gpt-5.5") }),
+    request: (input) => {
+      const headers = new Headers(input.headers)
+      headers.set("x-seen-url", input.url)
+      return Effect.succeed(new Response(null, { headers }))
+    },
+  })
+  const response = yield* event.request(new Request(url, { method: "POST", body: "{}" }))
+  return { url: response.headers.get("x-seen-url"), headers: Object.fromEntries(response.headers.entries()) }
+})
+
 describe("OpenAIPlugin", () => {
   it.effect("registers browser and headless ChatGPT OAuth methods", () =>
     Effect.gen(function* () {
@@ -100,33 +115,9 @@ describe("OpenAIPlugin", () => {
       })
       yield* addPlugin()
 
-      const request = yield* (yield* PluginHooks.Service).trigger("session", "request", {
-        sessionID: Session.ID.make("ses_test"),
-        agent: Agent.ID.make("build"),
-        model: Model.Ref.make({ providerID: Provider.ID.openai, id: Model.ID.make("gpt-5.5") }),
-        url: "https://api.openai.com/v1/responses",
-        method: "POST",
-        headers: {},
-        body: "{}",
-      })
-      const custom = yield* (yield* PluginHooks.Service).trigger("session", "request", {
-        sessionID: Session.ID.make("ses_test"),
-        agent: Agent.ID.make("build"),
-        model: Model.Ref.make({ providerID: Provider.ID.make("custom-openai"), id: Model.ID.make("gpt-5.5") }),
-        url: "https://custom.example/v1/responses",
-        method: "POST",
-        headers: {},
-        body: "{}",
-      })
-      const proxy = yield* (yield* PluginHooks.Service).trigger("session", "request", {
-        sessionID: Session.ID.make("ses_test"),
-        agent: Agent.ID.make("build"),
-        model: Model.Ref.make({ providerID: Provider.ID.openai, id: Model.ID.make("gpt-5.5") }),
-        url: "https://proxy.example/v1/responses?region=us",
-        method: "POST",
-        headers: {},
-        body: "{}",
-      })
+      const request = yield* http(Provider.ID.openai, "https://api.openai.com/v1/responses")
+      const custom = yield* http(Provider.ID.make("custom-openai"), "https://custom.example/v1/responses")
+      const proxy = yield* http(Provider.ID.openai, "https://proxy.example/v1/responses?region=us")
 
       const provider = required(yield* catalog.provider.get(Provider.ID.openai))
       expect(provider.package).toBe("@opencode-ai/ai/providers/openai")
@@ -134,7 +125,7 @@ describe("OpenAIPlugin", () => {
       expect(provider.headers).toMatchObject({ "chatgpt-account-id": "acct_123" })
       expect(request.url).toBe("https://chatgpt.com/backend-api/codex/responses")
       expect(request.headers).toMatchObject({ originator: "opencode", "session-id": "ses_test" })
-      expect(custom.headers).toEqual({})
+      expect(custom.headers).not.toHaveProperty("originator")
       expect(proxy.url).toBe("https://proxy.example/v1/responses?region=us")
       expect(proxy.headers).toMatchObject({ originator: "opencode", "session-id": "ses_test" })
       const eligible = required(yield* catalog.model.get(Provider.ID.openai, Model.ID.make("gpt-5.5")))
@@ -184,21 +175,13 @@ describe("OpenAIPlugin", () => {
       })
       yield* addPlugin()
 
-      const request = yield* (yield* PluginHooks.Service).trigger("session", "request", {
-        sessionID: Session.ID.make("ses_test"),
-        agent: Agent.ID.make("build"),
-        model: Model.Ref.make({ providerID: Provider.ID.openai, id: Model.ID.make("gpt-5.5") }),
-        url: "https://api.openai.com/v1/responses",
-        method: "POST",
-        headers: {},
-        body: "{}",
-      })
+      const request = yield* http(Provider.ID.openai, "https://api.openai.com/v1/responses")
 
       const model = required(yield* catalog.model.get(Provider.ID.openai, Model.ID.make("gpt-5.5")))
       expect(model.package).toBe("@opencode-ai/ai/providers/openai")
       expect(model.enabled).toBe(true)
       expect(model.limit).toEqual({ context: 1_050_000, input: 922_000, output: 128_000 })
-      expect(request.headers).toEqual({})
+      expect(request.headers).not.toHaveProperty("originator")
       expect(required(yield* catalog.model.get(Provider.ID.openai, Model.ID.make("gpt-4.1"))).enabled).toBe(true)
     }),
   )

packages/core/test/session-runner-recorded.test.ts

@@ -1,4 +1,5 @@
 import { HttpRecorder } from "@opencode-ai/http-recorder"
+import type { SessionHookRegistration } from "@opencode-ai/plugin/effect/session"
 import * as OpenAIChat from "@opencode-ai/ai/protocols/openai-chat"
 import { Auth, LLMClient, RequestExecutor } from "@opencode-ai/ai/route"
 import { Catalog } from "@opencode-ai/core/catalog"
@@ -41,6 +42,7 @@ import { SystemPromptPlugin } from "@opencode-ai/core/plugin/system-prompt"
 import { describe, expect } from "bun:test"
 import { eq } from "drizzle-orm"
 import { Effect, Layer, Stream } from "effect"
+import { HttpClient, HttpClientResponse } from "effect/unstable/http"
 import path from "node:path"
 import { testEffect } from "./lib/effect"
 import { agentHost, catalogHost, host } from "./plugin/host"
@@ -104,37 +106,39 @@ const promptCatalog = Layer.mock(Catalog.Service, {
     small: () => Effect.succeed(undefined),
   },
 })
-const runnerLayer = AppNodeBuilder.build(SessionRunnerLLM.node, [
-  [Snapshot.node, Snapshot.noopLayer],
-  [LayerNodePlatform.llmClient, client],
-  [SessionRunnerModel.node, models],
-  [InstructionBuiltIns.node, systemContext],
-  [InstructionDiscovery.node, instructionContext],
-  [Location.node, Location.boundNode({ directory: AbsolutePath.make("/project") })],
-  [SkillInstructions.node, skillInstructions],
-  [ReferenceInstructions.node, referenceInstructions],
-  [McpInstructions.node, mcpInstructions],
-  [Config.node, config],
-  [Permission.node, permission],
-  [PluginSupervisor.node, pluginSupervisor],
-])
-const execution = Layer.effect(
-  SessionExecution.Service,
-  Effect.gen(function* () {
-    const sessionRunner = yield* SessionRunner.Service
-    const coordinator = yield* SessionRunCoordinator.make<Session.ID, SessionRunner.RunError>({
-      drain: (sessionID, force) => sessionRunner.drain({ sessionID, force }),
-    })
-    return SessionExecution.Service.of({
-      active: coordinator.active,
-      resume: coordinator.run,
-      wake: coordinator.wake,
-      interrupt: coordinator.interrupt,
-      awaitIdle: coordinator.awaitIdle,
-    })
-  }),
-).pipe(Layer.provide(runnerLayer))
-const it = testEffect(
+const runnerLayer = (llmClient: Layer.Layer<typeof LLMClient.Service>) =>
+  AppNodeBuilder.build(SessionRunnerLLM.node, [
+    [Snapshot.node, Snapshot.noopLayer],
+    [LayerNodePlatform.llmClient, llmClient],
+    [SessionRunnerModel.node, models],
+    [InstructionBuiltIns.node, systemContext],
+    [InstructionDiscovery.node, instructionContext],
+    [Location.node, Location.boundNode({ directory: AbsolutePath.make("/project") })],
+    [SkillInstructions.node, skillInstructions],
+    [ReferenceInstructions.node, referenceInstructions],
+    [McpInstructions.node, mcpInstructions],
+    [Config.node, config],
+    [Permission.node, permission],
+    [PluginSupervisor.node, pluginSupervisor],
+  ])
+const execution = (llmClient: Layer.Layer<typeof LLMClient.Service>) =>
+  Layer.effect(
+    SessionExecution.Service,
+    Effect.gen(function* () {
+      const sessionRunner = yield* SessionRunner.Service
+      const coordinator = yield* SessionRunCoordinator.make<Session.ID, SessionRunner.RunError>({
+        drain: (sessionID, force) => sessionRunner.drain({ sessionID, force }),
+      })
+      return SessionExecution.Service.of({
+        active: coordinator.active,
+        resume: coordinator.run,
+        wake: coordinator.wake,
+        interrupt: coordinator.interrupt,
+        awaitIdle: coordinator.awaitIdle,
+      })
+    }),
+  ).pipe(Layer.provide(runnerLayer(llmClient)))
+const testLayer = (llmClient: Layer.Layer<typeof LLMClient.Service>) =>
   AppNodeBuilder.build(
     LayerNode.group([
       Database.node,
@@ -156,7 +160,7 @@ const it = testEffect(
       Session.node,
     ]),
     [
-      [LayerNodePlatform.llmClient, client],
+      [LayerNodePlatform.llmClient, llmClient],
       [Permission.node, permission],
       [Catalog.node, promptCatalog],
       [SessionRunnerModel.node, models],
@@ -168,10 +172,10 @@ const it = testEffect(
       [Config.node, config],
       [Snapshot.node, Snapshot.noopLayer],
       [PluginSupervisor.node, pluginSupervisor],
-      [SessionExecution.node, execution],
+      [SessionExecution.node, execution(llmClient)],
     ],
-  ),
-)
+  )
+const it = testEffect(testLayer(client))
 const sessionID = Session.ID.make("ses_runner_recorded")
 
 describe("SessionRunnerLLM recorded", () => {
@@ -189,7 +193,12 @@ describe("SessionRunnerLLM recorded", () => {
       const pluginHost = host({
         agent: agentHost(agents),
         catalog: catalogHost(catalog),
-        session: { hook: (name, callback) => hooks.register("session", name, callback) },
+        session: {
+          hook: (...registration: SessionHookRegistration) => {
+            if (registration[0] === "http") return Effect.die("unused session HTTP hook")
+            return hooks.register("session", ...registration)
+          },
+        },
       })
       yield* Effect.forEach(SystemPromptPlugin.Plugins, (plugin) => plugin.effect(pluginHost), { discard: true })
       const { db } = yield* Database.Service
@@ -247,3 +256,99 @@ describe("SessionRunnerLLM recorded", () => {
     }),
   )
 })
+
+describe("SessionModelRequest HTTP bridge", () => {
+  const bodies: Uint8Array[] = []
+  const methods: string[] = []
+  const response = [
+    'data: {"id":"chatcmpl_test","object":"chat.completion.chunk","created":0,"model":"gpt-4o-mini","choices":[{"index":0,"delta":{"role":"assistant","content":"Hello!"},"finish_reason":null}]}',
+    'data: {"id":"chatcmpl_test","object":"chat.completion.chunk","created":0,"model":"gpt-4o-mini","choices":[{"index":0,"delta":{},"finish_reason":"stop"}]}',
+    "data: [DONE]",
+    "",
+  ].join("\n\n")
+  const transport = Layer.succeed(
+    HttpClient.HttpClient,
+    HttpClient.make((request) =>
+      Effect.sync(() => {
+        if (request.body._tag !== "Uint8Array") throw new Error(`Unexpected request body: ${request.body._tag}`)
+        methods.push(request.method)
+        bodies.push(request.body.body.slice())
+        return HttpClientResponse.fromWeb(
+          request,
+          new Response(response, { headers: { "content-type": "text/event-stream" } }),
+        )
+      }),
+    ),
+  )
+  const retryIt = testEffect(
+    testLayer(LLMClient.layer.pipe(Layer.provide(RequestExecutor.layer.pipe(Layer.provide(transport))))),
+  )
+
+  retryIt.effect("lets an Effect plugin send the same POST Request twice", () =>
+    Effect.gen(function* () {
+      bodies.length = 0
+      methods.length = 0
+      const agents = yield* Agent.Service
+      const catalog = yield* Catalog.Service
+      const hooks = yield* PluginHooks.Service
+      yield* agents.transform((draft) =>
+        draft.update(Agent.ID.make("build"), (agent) => {
+          agent.mode = "primary"
+          agent.permissions.push({ action: "execute", resource: "*", effect: "deny" })
+        }),
+      )
+      const pluginHost = host({
+        agent: agentHost(agents),
+        catalog: catalogHost(catalog),
+        session: {
+          hook: (...registration: SessionHookRegistration) => {
+            if (registration[0] !== "http") return hooks.register("session", ...registration)
+            const middleware = registration[1]
+            return hooks.register("session", "http", (event) =>
+              Effect.sync(() => {
+                const next = event.request
+                event.request = (request) => middleware(event, request, next)
+              }),
+            )
+          },
+        },
+      })
+      yield* pluginHost.session.hook("http", (_context, request, next) =>
+        Effect.gen(function* () {
+          yield* next(request).pipe(Effect.flatMap((response) => Effect.promise(() => response.text())))
+          return yield* next(request)
+        }),
+      )
+      yield* Effect.forEach(SystemPromptPlugin.Plugins, (plugin) => plugin.effect(pluginHost), { discard: true })
+      const { db } = yield* Database.Service
+      yield* db
+        .insert(ProjectTable)
+        .values({ id: Project.ID.global, worktree: AbsolutePath.make("/project"), sandboxes: [] })
+        .onConflictDoNothing()
+        .run()
+        .pipe(Effect.orDie)
+      const retrySessionID = Session.ID.make("ses_model_request_http_retry")
+      yield* db
+        .insert(SessionTable)
+        .values({
+          id: retrySessionID,
+          project_id: Project.ID.global,
+          slug: "test",
+          directory: "/project",
+          title: "test",
+          version: "test",
+        })
+        .run()
+        .pipe(Effect.orDie)
+      const session = yield* Session.Service
+      yield* session.prompt({ sessionID: retrySessionID, text: "Say hello.", resume: false })
+
+      yield* session.resume(retrySessionID)
+
+      expect(methods).toEqual(["POST", "POST"])
+      expect(bodies).toHaveLength(2)
+      expect(bodies[0]?.byteLength).toBeGreaterThan(0)
+      expect(bodies[1]).toEqual(bodies[0])
+    }),
+  )
+})

packages/core/test/session-runner.test.ts

@@ -14,6 +14,7 @@ import {
 } from "@opencode-ai/ai"
 import * as OpenAIChat from "@opencode-ai/ai/protocols/openai-chat"
 import { TestLLM } from "@opencode-ai/ai/testing"
+import type { SessionHookRegistration } from "@opencode-ai/plugin/effect/session"
 import { Catalog } from "@opencode-ai/core/catalog"
 import { Database } from "@opencode-ai/core/database/database"
 import { makeLocationNode } from "@opencode-ai/util/effect/app-node"
@@ -474,7 +475,12 @@ const setup = Effect.gen(function* () {
   const pluginHost = host({
     agent: agentHost(agents),
     catalog: catalogHost(catalog),
-    session: { hook: (name, callback) => hooks.register("session", name, callback) },
+    session: {
+      hook: (...registration: SessionHookRegistration) => {
+        if (registration[0] === "http") return Effect.die("unused session HTTP hook")
+        return hooks.register("session", ...registration)
+      },
+    },
   })
   yield* Effect.forEach(SystemPromptPlugin.Plugins, (plugin) => plugin.effect(pluginHost), {
     discard: true,

packages/core/test/shared-schema.test.ts

@@ -166,7 +166,7 @@ test("Core reuses the canonical shared schemas", async () => {
   ]
   for (const [core, shared] of schemas) expect(core).toBe(shared)
 
-  expect(Agent.Info.empty(Agent.ID.make("test"))).toEqual(Agent.Info.empty(Agent.ID.make("test")))
+  expect(Agent.Info.default(Agent.ID.make("test"))).toEqual(Agent.Info.default(Agent.ID.make("test")))
   expect(coreModel.Info.default(coreProvider.ID.make("test"), coreModel.ID.make("model"))).toEqual(
     Model.Info.default(Provider.ID.make("test"), Model.ID.make("model")),
   )

packages/core/test/skill/instructions.test.ts

@@ -47,7 +47,7 @@ const layer = (list: () => Skill.Info[]) =>
 describe("SkillInstructions", () => {
   it.effect("renders described agent skills and updates the complete available list", () => {
     const agent = Agent.Info.make({
-      ...Agent.Info.empty(build),
+      ...Agent.Info.default(build),
       permissions: [{ action: "skill", resource: "denied", effect: "deny" }],
     })
     let skills = [hidden, denied, manual, effect]
@@ -80,7 +80,7 @@ describe("SkillInstructions", () => {
   })
 
   it.effect("announces added and removed skills as deltas without restating the list", () => {
-    const agent = Agent.Info.make(Agent.Info.empty(build))
+    const agent = Agent.Info.make(Agent.Info.default(build))
     const debugging = Skill.Info.make({
       id: Skill.ID.make("debugging"),
       name: Skill.Name.make("Debugging"),
@@ -117,7 +117,7 @@ describe("SkillInstructions", () => {
   })
 
   it.effect("restates the full skill list when a description changes", () => {
-    const agent = Agent.Info.make(Agent.Info.empty(build))
+    const agent = Agent.Info.make(Agent.Info.default(build))
     let skills = [effect]
     return Effect.gen(function* () {
       const instructions = yield* SkillInstructions.Service
@@ -138,7 +138,7 @@ describe("SkillInstructions", () => {
 
   it.effect("omits instructions when the selected agent denies all skills", () => {
     const agent = Agent.Info.make({
-      ...Agent.Info.empty(build),
+      ...Agent.Info.default(build),
       permissions: [{ action: "skill", resource: "*", effect: "deny" }],
     })
     return Effect.gen(function* () {
@@ -149,7 +149,7 @@ describe("SkillInstructions", () => {
 
   it.effect("omits instructions when a resource-specific denial follows the global denial", () => {
     const agent = Agent.Info.make({
-      ...Agent.Info.empty(build),
+      ...Agent.Info.default(build),
       permissions: [
         { action: "skill", resource: "*", effect: "deny" },
         { action: "skill", resource: "hidden", effect: "deny" },
@@ -163,7 +163,7 @@ describe("SkillInstructions", () => {
 
   it.effect("retains specifically allowed skills after a global denial", () => {
     const agent = Agent.Info.make({
-      ...Agent.Info.empty(build),
+      ...Agent.Info.default(build),
       permissions: [
         { action: "skill", resource: "*", effect: "deny" },
         { action: "skill", resource: "effect", effect: "allow" },
@@ -179,7 +179,7 @@ describe("SkillInstructions", () => {
 
   it.effect("omits instructions when a specifically allowed skill is denied again", () => {
     const agent = Agent.Info.make({
-      ...Agent.Info.empty(build),
+      ...Agent.Info.default(build),
       permissions: [
         { action: "skill", resource: "*", effect: "deny" },
         { action: "skill", resource: "effect", effect: "allow" },

packages/plugin/src/effect/session.ts

@@ -1,11 +1,10 @@
 import type { SessionApi } from "@opencode-ai/client/effect/api"
 import type { Message, SystemPart } from "@opencode-ai/ai"
-import type { HttpRequest } from "@opencode-ai/ai/route"
 import type { Agent } from "@opencode-ai/schema/agent"
 import type { Model } from "@opencode-ai/schema/model"
 import type { Session } from "@opencode-ai/schema/session"
-import type { JsonSchema } from "effect"
-import type { Hooks } from "./registration.js"
+import type { Effect, JsonSchema, Scope } from "effect"
+import type { Registration } from "./registration.js"
 
 export interface SessionContext {
   readonly sessionID: Session.ID
@@ -16,20 +15,39 @@ export interface SessionContext {
   tools: Record<string, { description: string; input: JsonSchema.JsonSchema }>
 }
 
-export interface SessionRequest extends HttpRequest {
+export interface SessionHttpContext {
   readonly sessionID: Session.ID
   readonly agent: Agent.ID
   readonly model: Model.Ref
 }
 
+export type SessionHttpMiddleware = (
+  context: SessionHttpContext,
+  request: Request,
+  next: (request: Request) => Effect.Effect<Response, Error>,
+) => Effect.Effect<Response, Error>
+
 export interface SessionHooks {
   readonly context: SessionContext
-  readonly request: SessionRequest
+}
+
+export type SessionHookRegistration =
+  | {
+      [Name in keyof SessionHooks]: [name: Name, callback: (event: SessionHooks[Name]) => Effect.Effect<void>]
+    }[keyof SessionHooks]
+  | [name: "http", middleware: SessionHttpMiddleware]
+
+export interface SessionHook {
+  <Name extends keyof SessionHooks>(
+    name: Name,
+    callback: (event: SessionHooks[Name]) => Effect.Effect<void>,
+  ): Effect.Effect<Registration, never, Scope.Scope>
+  (name: "http", middleware: SessionHttpMiddleware): Effect.Effect<Registration, never, Scope.Scope>
 }
 
 export type SessionDomain = Pick<
   SessionApi<unknown>,
   "create" | "get" | "prompt" | "generate" | "command" | "synthetic" | "interrupt" | "rename" | "wait"
 > & {
-  readonly hook: Hooks<SessionHooks>
+  readonly hook: SessionHook
 }

packages/plugin/src/promise/session.ts

@@ -1,11 +1,10 @@
 import type { SessionApi } from "@opencode-ai/client/promise/api"
 import type { Message, SystemPart } from "@opencode-ai/ai"
-import type { HttpRequest } from "@opencode-ai/ai/route"
 import type { Agent } from "@opencode-ai/schema/agent"
 import type { Model } from "@opencode-ai/schema/model"
 import type { Session } from "@opencode-ai/schema/session"
 import type { JsonSchema } from "effect"
-import type { Hooks } from "./registration.js"
+import type { Registration } from "./registration.js"
 
 export interface SessionContext {
   readonly sessionID: Session.ID
@@ -16,20 +15,39 @@ export interface SessionContext {
   tools: Record<string, { description: string; input: JsonSchema.JsonSchema }>
 }
 
-export interface SessionRequest extends HttpRequest {
+export interface SessionHttpContext {
   readonly sessionID: Session.ID
   readonly agent: Agent.ID
   readonly model: Model.Ref
 }
 
+export type SessionHttpMiddleware = (
+  context: SessionHttpContext,
+  request: Request,
+  next: (request: Request) => Promise<Response>,
+) => Promise<Response> | Response
+
 export interface SessionHooks {
   readonly context: SessionContext
-  readonly request: SessionRequest
+}
+
+export type SessionHookRegistration =
+  | {
+      [Name in keyof SessionHooks]: [name: Name, callback: (event: SessionHooks[Name]) => Promise<void> | void]
+    }[keyof SessionHooks]
+  | [name: "http", middleware: SessionHttpMiddleware]
+
+export interface SessionHook {
+  <Name extends keyof SessionHooks>(
+    name: Name,
+    callback: (event: SessionHooks[Name]) => Promise<void> | void,
+  ): Promise<Registration>
+  (name: "http", middleware: SessionHttpMiddleware): Promise<Registration>
 }
 
 export type SessionDomain = Pick<
   SessionApi,
   "create" | "get" | "prompt" | "generate" | "command" | "synthetic" | "interrupt"
 > & {
-  readonly hook: Hooks<SessionHooks>
+  readonly hook: SessionHook
 }

packages/schema/src/agent.ts

@@ -36,7 +36,7 @@ export const Info = Schema.Struct({
   .annotate({ identifier: "Agent.Info" })
   .pipe(
     statics(() => ({
-      empty: (id: ID) =>
+      default: (id: ID) =>
         ({
           id,
           name: Name.make(id),

packages/www/content/docs/build/plugins.mdx

@@ -239,16 +239,28 @@ without restarting OpenCode.
 
 ### Runtime hooks
 
-Runtime hooks intercept live operations. Their event objects expose specific
-mutable fields:
+Runtime hooks intercept live operations:
 
 | Hook                                        | Mutable fields                                                                 |
 | ------------------------------------------- | ------------------------------------------------------------------------------ |
 | `ctx.aisdk.hook("sdk", callback)`           | `sdk`, after inspecting `model`, `package`, and `options`                      |
 | `ctx.aisdk.hook("language", callback)`      | `language`, after inspecting `model`, `sdk`, and `options`                     |
-| `ctx.session.hook("request", callback)`     | `system`, `messages`, and the `tools` record immediately before model dispatch |
+| `ctx.session.hook("context", callback)`     | `system`, `messages`, and the `tools` record immediately before model dispatch |
+| `ctx.session.hook("http", middleware)`     | The model's HTTP request and response                                          |
 | `ctx.tool.hook("execute.before", callback)` | `input`, before the selected tool executes                                     |
-| `ctx.tool.hook("execute.after", callback)`  | Terminal `result` on success or `error` on failure                            |
+| `ctx.tool.hook("execute.after", callback)`  | Terminal `result` on success or `error` on failure                             |
+
+HTTP middleware can modify requests, inspect responses, retry, or return a
+response without calling the provider. It applies to native models; AI SDK
+models do not currently pass through this hook.
+
+```ts
+await ctx.session.hook("http", async (event, request, next) => {
+  request.headers.set("x-session-id", event.sessionID)
+  const response = await next(request)
+  return response
+})
+```
 
 For example, remove a tool from selected model requests and normalize another
 tool's input:
@@ -259,7 +271,7 @@ import { Plugin } from "@opencode-ai/plugin"
 export default Plugin.define({
   id: "acme.guards",
   setup: async (ctx) => {
-    await ctx.session.hook("request", (event) => {
+    await ctx.session.hook("context", (event) => {
       delete event.tools.write
     })