fix(core): settle interrupted assistant steps (#33266)

Author: kitlangton

packages/core/src/session/runner/llm.ts

@@ -305,14 +305,7 @@ export const layer = Layer.effect(
           const llmFailure = failure instanceof LLMError ? failure : undefined
           if (llmFailure && !publisher.hasProviderError()) {
             yield* withPublication(publisher.failUnsettledTools("Provider did not return a tool result", true))
-            yield* withPublication(
-              events.publish(SessionEvent.Step.Failed, {
-                sessionID: session.id,
-                timestamp: yield* DateTime.now,
-                assistantMessageID: yield* publisher.startAssistant(),
-                error: { type: "unknown", message: llmFailure.reason.message },
-              }),
-            )
+            yield* withPublication(publisher.failAssistant(llmFailure.reason.message))
           }
           if (stream._tag === "Failure" && Cause.hasInterrupts(stream.cause)) yield* FiberSet.clear(toolFibers)
           const settled = yield* restore(awaitToolFibers(toolFibers)).pipe(Effect.exit)
@@ -327,6 +320,8 @@ export const layer = Layer.effect(
           ) {
             yield* FiberSet.clear(toolFibers)
             yield* withPublication(publisher.failUnsettledTools("Tool execution interrupted"))
+            if (publisher.hasActiveAssistant())
+              yield* withPublication(publisher.failAssistant("Provider turn interrupted"))
           }
           if (settled._tag === "Failure" && !Cause.hasInterrupts(settled.cause)) {
             const failure = Cause.squash(settled.cause)

packages/core/src/session/runner/publish-llm-event.ts

@@ -65,11 +65,14 @@ export const createLLMEventPublisher = (events: EventV2.Interface, input: Input)
   >()
   const timestamp = DateTime.now
   let assistantMessageID: SessionMessage.ID | undefined
+  let assistantActive = false
+  let assistantFailed = false
   let providerFailed = false
 
   const startAssistant = Effect.fnUntraced(function* () {
     if (assistantMessageID !== undefined) return assistantMessageID
     assistantMessageID = SessionMessage.ID.create()
+    assistantActive = true
     yield* events.publish(SessionEvent.Step.Started, {
       ...input,
       assistantMessageID,
@@ -190,6 +193,20 @@ export const createLLMEventPublisher = (events: EventV2.Interface, input: Input)
     yield* flushFragments()
   })
 
+  const failAssistant = Effect.fnUntraced(function* (message: string) {
+    if (assistantFailed) return
+    yield* flush()
+    const assistantMessageID = yield* startAssistant()
+    assistantActive = false
+    assistantFailed = true
+    yield* events.publish(SessionEvent.Step.Failed, {
+      sessionID: input.sessionID,
+      timestamp: yield* timestamp,
+      assistantMessageID,
+      error: { type: "unknown", message },
+    })
+  })
+
   const failUnsettledTools = Effect.fn("SessionRunner.failUnsettledTools")(function* (
     message: string,
     hostedOnly = false,
@@ -375,6 +392,7 @@ export const createLLMEventPublisher = (events: EventV2.Interface, input: Input)
       }
       case "step-finish":
         yield* flush()
+        assistantActive = false
         yield* events.publish(SessionEvent.Step.Ended, {
           sessionID: input.sessionID,
           timestamp: yield* timestamp,
@@ -388,24 +406,19 @@ export const createLLMEventPublisher = (events: EventV2.Interface, input: Input)
         return
       case "provider-error":
         providerFailed = true
-        yield* flush()
-        yield* events.publish(SessionEvent.Step.Failed, {
-          sessionID: input.sessionID,
-          timestamp: yield* timestamp,
-          assistantMessageID: yield* startAssistant(),
-          error: { type: "unknown", message: event.message },
-        })
+        yield* failAssistant(event.message)
         return
     }
   })
 
   return {
     publish,
     flush,
+    failAssistant,
     failUnsettledTools,
+    hasActiveAssistant: () => assistantActive,
     hasAssistantStarted: () => assistantMessageID !== undefined,
     hasProviderError: () => providerFailed,
-    startAssistant,
     assistantMessageID: assistantMessageIDForTool,
   }
 }

packages/core/src/session/runner/to-llm-message.ts

@@ -82,12 +82,21 @@ const assistant = (message: SessionMessage.Assistant, model: Model) => {
     const result = toolResult(item, sameModel ? (item.provider?.resultMetadata ?? item.provider?.metadata) : undefined)
     return item.provider?.executed === true && result ? [call, result] : [call]
   })
+  const meaningful = content.filter((part) => {
+    if (part.type === "text") return part.text !== ""
+    if (part.type !== "reasoning") return true
+    return part.text !== "" || (part.providerMetadata !== undefined && Object.keys(part.providerMetadata).length > 0)
+  })
   const results = message.content
     .filter((item): item is SessionMessage.AssistantTool => item.type === "tool" && item.provider?.executed !== true)
     .map((item) => toolResult(item, sameModel ? (item.provider?.resultMetadata ?? item.provider?.metadata) : undefined))
     .filter((message) => message !== undefined)
     .map(Message.tool)
-  return [Message.make({ id: message.id, role: "assistant", content, metadata: message.metadata }), ...results]
+  if (meaningful.length === 0) return results
+  return [
+    Message.make({ id: message.id, role: "assistant", content: meaningful, metadata: message.metadata }),
+    ...results,
+  ]
 }
 
 function toLLMMessage(message: SessionMessage.Message, model: Model): Message[] {

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

@@ -14,6 +14,39 @@ const id = (value: string) => SessionMessage.ID.make(`msg_${value}`)
 const model = Model.make({ id: "model", provider: "provider", route: OpenAIChat.route })
 
 describe("toLLMMessages", () => {
+  test("omits empty assistant turns", () => {
+    const assistant = (value: string, content: SessionMessage.Assistant["content"]) =>
+      new SessionMessage.Assistant({
+        id: id(value),
+        type: "assistant",
+        agent: "build",
+        model: { id: ModelV2.ID.make("model"), providerID: ProviderV2.ID.make("provider") },
+        content,
+        time: { created, completed: created },
+      })
+    const messages = toLLMMessages(
+      [
+        assistant("empty", []),
+        assistant("empty-text", [new SessionMessage.AssistantText({ type: "text", id: "empty", text: "" })]),
+        assistant("empty-reasoning", [
+          new SessionMessage.AssistantReasoning({ type: "reasoning", id: "empty-reasoning", text: "" }),
+        ]),
+        assistant("text", [new SessionMessage.AssistantText({ type: "text", id: "text", text: "Partial" })]),
+        assistant("reasoning", [
+          new SessionMessage.AssistantReasoning({
+            type: "reasoning",
+            id: "reasoning",
+            text: "",
+            providerMetadata: { anthropic: { signature: "sig_1" } },
+          }),
+        ]),
+      ],
+      model,
+    )
+
+    expect(messages.map((message) => message.id)).toEqual([id("text"), id("reasoning")])
+  })
+
   test("maps every top-level V2 Session message type", () => {
     const file = new FileAttachment({ uri: "data:image/png;base64,aGVsbG8=", mime: "image/png", name: "hello.png" })
     const messages = toLLMMessages(

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

@@ -547,6 +547,8 @@ const verifyPartialFlushOnInterruption = (kind: FragmentKind) =>
       { type: "user", text: prompt },
       {
         type: "assistant",
+        finish: "error",
+        error: { type: "unknown", message: "Provider turn interrupted" },
         content: [
           kind === "tool input"
             ? { type: "tool", id: fragmentID(kind, "interrupted"), state: { status: "error" } }