fix(spawn): capture tool args from tool_execution_start + journal filesTouched (#60)

src/engine/spawnSubagent.ts

@@ -378,6 +378,9 @@ export async function spawnSubagent(opts: SpawnOptions): Promise<SpawnResult> {
     // #61: executed-tool count — a "completed" run with zero tool calls is the premature-return
     // shape (read/narrate/end without acting); the tool flags it so the controller verifies.
     let toolCallCount = 0;
+    // #60: args live on tool_execution_start (the SDK's end event carries none) — capture per
+    // toolCallId so extractTouchedFiles + the journal see the real args on real runs.
+    const pendingToolArgs = new Map<string, unknown>();
 
     // #23: liveness — classify events into a short, content-free class string for the widget.
     // Names the tool (safe — tool name is not args/result) so the operator sees "what's happening"
@@ -445,13 +448,20 @@ export async function spawnSubagent(opts: SpawnOptions): Promise<SpawnResult> {
           aborted = true;
           void session.abort();
         }
+      } else if (e.type === "tool_execution_start") {
+        pendingToolArgs.set((e as { toolCallId?: string }).toolCallId ?? "", (e as { args?: unknown }).args);   // #60
       } else if (e.type === "tool_execution_end") {
         // #49: track mutated files for the structured partial-result report.
         toolCallCount += 1;   // #61
+        const toolCallId = (e as { toolCallId?: string }).toolCallId ?? "";
+        // #60: real args from the start event; `?? e.args` keeps hand-rolled fakes that put args
+        // on the end event working (the SDK never does).
+        const args = pendingToolArgs.get(toolCallId) ?? (e as { args?: unknown }).args;
+        pendingToolArgs.delete(toolCallId);
         sawAssistantAfterLastTool = false;   // cut mid-tool-work unless a message follows
-        for (const f of extractTouchedFiles((e as { toolName?: string }).toolName ?? "", (e as { args?: unknown }).args)) filesTouched.add(f);
+        for (const f of extractTouchedFiles((e as { toolName?: string }).toolName ?? "", args)) filesTouched.add(f);
         try {
-          opts.runLog?.append(runId, buildToolEvent((e as any).toolName, (e as any).args, (e as any).result, (e as any).isError ?? false, turnIdx));
+          opts.runLog?.append(runId, buildToolEvent((e as any).toolName, args, (e as any).result, (e as any).isError ?? false, turnIdx));
         } catch { /* best-effort */ }
       }
       // #23: liveness heartbeat — update the run record on meaningful events so the fleet widget
@@ -595,6 +605,8 @@ async function finishRun(
       error,
       // #61: executed-tool count (the zero-work premature-return signal, post-hoc too).
       toolCallCount,
+      // #60: what the run mutated (was only on the SpawnResult; the durable journal lacked it).
+      filesTouched,
     });
   } catch { /* best-effort: journal is the index, not the product */ }
   // SPEC-4: lifecycle phase children skip the per-run todo reconciliation — the lifecycle

src/runtime/run-log.ts

@@ -37,6 +37,8 @@ export interface RunEndedEvent {
   error?: string;
   /** #61: executed-tool count (the zero-work premature-return signal, post-hoc too). */
   toolCallCount?: number;
+  /** #60: file paths the run mutated (post-hoc — was previously SpawnResult-only). */
+  filesTouched?: string[];
 }
 export type RunLogEvent = RunMetaEvent | MessageEvent | ToolEvent | RunEndedEvent;
 

test/spawn-subagent-runlog.test.mts

@@ -270,3 +270,65 @@ test("#61 follow-up: claude tool_use input feeds filesTouched (Edit block path e
   });
   deepStrictEqual(res.filesTouched, ["/repo/src/a.ts"], "claude Edit blocks contribute to filesTouched (#49 parity)");
 });
+
+test("#60: REAL SDK shape — args live on tool_execution_start (end has none); capture restores filesTouched + journal args", async () => {
+  // The pi SDK emits tool_execution_end WITHOUT args ({toolCallId, toolName, result, isError});
+  // only tool_execution_start carries args. The old code read args off the END event (always
+  // undefined on real runs) — filesTouched stayed empty and the journal serialized args as "".
+  const handlers: Array<(e: any) => void> = [];
+  const realShapeChild: ChildSession = {
+    prompt: async () => {
+      for (const h of handlers) h({ type: "session_init", backendSessionId: "s60" });
+      for (const h of handlers) h({ type: "turn_start", turnIndex: 0 });
+      // exact pi SDK shape: start carries args, end does NOT
+      for (const h of handlers) h({ type: "tool_execution_start", toolCallId: "t1", toolName: "edit", args: { path: "/repo/src/real.ts", edits: [] } });
+      for (const h of handlers) h({ type: "tool_execution_end", toolCallId: "t1", toolName: "edit", result: "ok", isError: false });
+      for (const h of handlers) h({ type: "message_end", message: { role: "assistant", content: [{ type: "text", text: "done" }] } });
+      for (const h of handlers) h({ type: "turn_end", turnIndex: 0, message: {} as any, toolResults: [] });
+    },
+    subscribe: (h: any) => { handlers.push(h); return () => {}; },
+    abort: async () => {}, dispose: () => {},
+  };
+  const factory: ChildSessionFactory = { create: async () => ({ session: realShapeChild, model: "m" }) };
+  const log = new RunLog(logDir);
+  const h = harness(factory, log);
+  const res = await spawnSubagent({
+    agent: "g", task: "t", track: true, runLog: log,
+    registry: h.registry, todoSync: h.todoSync, runRegistry: h.runRegistry, lock: h.lock, backendRegistry: h.backendRegistry,
+    parentModel: PARENT, parentCwd: tmpDir,
+  });
+  strictEqual(res.status, "completed");
+  deepStrictEqual(res.filesTouched, ["/repo/src/real.ts"], "filesTouched populated from the START event's args");
+  const toolEvent = log.replay(res.runId).find((e) => e.type === "tool") as any;
+  ok(typeof toolEvent.args === "string" && toolEvent.args.includes("/repo/src/real.ts"), `journal tool event carries real args (was ""): ${toolEvent.args}`);
+});
+
+test("#60: filesTouched is journaled on run:ended when the turn budget cuts the run", async () => {
+  const handlers: Array<(e: any) => void> = [];
+  const budgetChild: ChildSession = {
+    prompt: async () => {
+      for (const h of handlers) h({ type: "session_init", backendSessionId: "s60b" });
+      for (const h of handlers) h({ type: "turn_start", turnIndex: 0 });
+      for (const h of handlers) h({ type: "tool_execution_start", toolCallId: "t1", toolName: "write", args: { path: "/repo/src/cut.ts" } });
+      for (const h of handlers) h({ type: "tool_execution_end", toolCallId: "t1", toolName: "write", result: "ok", isError: false });
+      for (const h of handlers) h({ type: "turn_end", turnIndex: 0, message: {} as any, toolResults: [] });
+      for (const h of handlers) h({ type: "turn_start", turnIndex: 1 });
+      for (const h of handlers) h({ type: "turn_end", turnIndex: 1, message: {} as any, toolResults: [] });
+    },
+    subscribe: (h: any) => { handlers.push(h); return () => {}; },
+    abort: async () => {}, dispose: () => {},
+  };
+  const factory: ChildSessionFactory = { create: async () => ({ session: budgetChild, model: "m" }) };
+  const log = new RunLog(logDir);
+  const h = harness(factory, log);
+  const res = await spawnSubagent({
+    agent: "g", task: "t", track: true, runLog: log, maxTurns: 1,
+    registry: h.registry, todoSync: h.todoSync, runRegistry: h.runRegistry, lock: h.lock, backendRegistry: h.backendRegistry,
+    parentModel: PARENT, parentCwd: tmpDir,
+  });
+  strictEqual(res.status, "failed", "turn budget exhausted");
+  ok((res.filesTouched ?? []).includes("/repo/src/cut.ts"), `SpawnResult.filesTouched: ${res.filesTouched}`);
+  const ended = log.replay(res.runId).at(-1) as any;
+  strictEqual(ended.type, "run:ended");
+  deepStrictEqual(ended.filesTouched, ["/repo/src/cut.ts"], "run:ended carries filesTouched (the #60 journal gap)");
+});