@zhixuan92/multi-model-agent 3.1.5 → 3.1.6

package/dist/http/execution-context.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"execution-context.d.ts","sourceRoot":"","sources":["../../src/http/execution-context.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,cAAc,EAAqB,MAAM,mCAAmC,CAAC;AAC3F,OAAO,KAAK,EAAE,gBAAgB,EAAyB,MAAM,mDAAmD,CAAC;AACjH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAErD;;;;;;;;;GASG;AACH,wBAAgB,qBAAqB,CACnC,IAAI,EAAE,WAAW,EACjB,EAAE,EAAE,cAAc,EAClB,OAAO,EAAE,MAAM,GACd,gBAAgB,CA0DlB"}
+{"version":3,"file":"execution-context.d.ts","sourceRoot":"","sources":["../../src/http/execution-context.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,cAAc,EAAqB,MAAM,mCAAmC,CAAC;AAC3F,OAAO,KAAK,EAAE,gBAAgB,EAAyB,MAAM,mDAAmD,CAAC;AACjH,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAErD;;;;;;;;;GASG;AACH,wBAAgB,qBAAqB,CACnC,IAAI,EAAE,WAAW,EACjB,EAAE,EAAE,cAAc,EAClB,OAAO,EAAE,MAAM,GACd,gBAAgB,CAiElB"}

package/dist/http/execution-context.js

@@ -21,15 +21,22 @@ export function buildExecutionContext(deps, pc, batchId) {
         // "running, 47s elapsed, worker: MiniMax-M2.7 (turn 2)" instead of
         // just "1/1 running". stageIndex is a reasonable turn proxy.
         entry.running = [{ worker: tick.provider, turn: Math.max(1, tick.stageIndex) }];
-        const headline = composeRunningHeadline({
-            tasksTotal: entry.tasksTotal ?? 1,
-            tasksStarted: entry.tasksStarted ?? 0,
-            tasksCompleted: entry.tasksCompleted ?? 0,
-            startedAt: entry.startedAt,
-            nowMs: Date.now(),
-            lastHeartbeatAt: entry.lastHeartbeatAt,
-            running: entry.running,
-        });
+        // Single-task batches get the rich per-stage headline composed by
+        // HeartbeatTimer (stage name, cost/ROI, file counts, tool calls).
+        // Multi-task batches fall back to the batch-level summary since no
+        // single worker's headline applies across tasks.
+        const tasksTotal = entry.tasksTotal ?? 1;
+        const headline = tasksTotal <= 1
+            ? tick.headline
+            : composeRunningHeadline({
+                tasksTotal,
+                tasksStarted: entry.tasksStarted ?? 0,
+                tasksCompleted: entry.tasksCompleted ?? 0,
+                startedAt: entry.startedAt,
+                nowMs: Date.now(),
+                lastHeartbeatAt: entry.lastHeartbeatAt,
+                running: entry.running,
+            });
         deps.batchRegistry.updateRunningHeadline(effectiveBatchId, headline);
         deps.logger.taskHeartbeat({
             batchId: effectiveBatchId,

package/dist/http/execution-context.js.map

@@ -1 +1 @@
-{"version":3,"file":"execution-context.js","sourceRoot":"","sources":["../../src/http/execution-context.ts"],"names":[],"mappings":"AAAA,gDAAgD;AAChD,OAAO,EAAE,cAAc,EAAE,sBAAsB,EAAE,MAAM,mCAAmC,CAAC;AAK3F;;;;;;;;;GASG;AACH,MAAM,UAAU,qBAAqB,CACnC,IAAiB,EACjB,EAAkB,EAClB,OAAe;IAEf,MAAM,eAAe,GAAG,CAAC,IAAuB,EAAE,EAAE;QAClD,MAAM,gBAAgB,GAAG,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC;QACjD,MAAM,KAAK,GAAG,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,gBAAgB,CAAC,CAAC;QACvD,IAAI,CAAC,KAAK;YAAE,OAAO;QACnB,KAAK,CAAC,eAAe,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QACnC,6DAA6D;QAC7D,mEAAmE;QACnE,6DAA6D;QAC7D,KAAK,CAAC,OAAO,GAAG,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,QAAQ,EAAE,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,UAAU,CAAC,EAAE,CAAC,CAAC;QAChF,MAAM,QAAQ,GAAG,sBAAsB,CAAC;YACtC,UAAU,EAAE,KAAK,CAAC,UAAU,IAAI,CAAC;YACjC,YAAY,EAAE,KAAK,CAAC,YAAY,IAAI,CAAC;YACrC,cAAc,EAAE,KAAK,CAAC,cAAc,IAAI,CAAC;YACzC,SAAS,EAAE,KAAK,CAAC,SAAS;YAC1B,KAAK,EAAE,IAAI,CAAC,GAAG,EAAE;YACjB,eAAe,EAAE,KAAK,CAAC,eAAe;YACtC,OAAO,EAAE,KAAK,CAAC,OAAO;SACvB,CAAC,CAAC;QACH,IAAI,CAAC,aAAa,CAAC,qBAAqB,CAAC,gBAAgB,EAAE,QAAQ,CAAC,CAAC;QACrE,IAAI,CAAC,MAAM,CAAC,aAAa,CAAC;YACxB,OAAO,EAAE,gBAAgB;YACzB,SAAS,EAAE,CAAC;YACZ,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,KAAK,EAAE,IAAI,CAAC,KAAK;SAClB,CAAC,CAAC;QACH,IAAI,IAAI,CAAC,WAAW,KAAK,SAAS,EAAE,CAAC;YACnC,IAAI,CAAC,MAAM,CAAC,eAAe,CAAC;gBAC1B,OAAO,EAAE,gBAAgB;gBACzB,SAAS,EAAE,CAAC;gBACZ,SAAS,EAAE,IAAI,CAAC,WAAW,CAAC,IAAI;gBAChC,OAAO,EAAE,IAAI,CAAC,WAAW,CAAC,EAAE;aAC7B,CAAC,CAAC;QACL,CAAC;IACH,CAAC,CAAC;IAEF,OAAO;QACL,cAAc,EAAE,EAAE;QAClB,MAAM,EAAE,IAAI,CAAC,MAAM;QACnB,MAAM,EAAE,IAAI,CAAC,MAAM;QACnB,iBAAiB,EAAE,EAAE,CAAC,aAAa;QACnC,eAAe,EAAE,CAAC,OAAe,EAAE,EAAE,CAAC,cAAc,CAAC,OAAiC,EAAE,IAAI,CAAC,MAAM,CAAC;QACpG,WAAW,EAAE,OAAO,CAAC,GAAG,CAAC,mBAAmB,CAAC;QAC7C,UAAU,EAAE,SAAS;QACrB,OAAO;QACP,eAAe;QACf,kBAAkB,EAAE,KAAK,EAAE,QAA+B,EAAE,EAAE;YAC5D,OAAO,IAAI,OAAO,CAA6B,CAAC,OAAO,EAAE,EAAE;gBACzD,MAAM,KAAK,GAAG,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;gBAC9C,IAAI,KAAK,EAAE,CAAC;oBACV,0EAA0E;oBAC1E,6CAA6C;oBAC7C,KAAK,CAAC,oBAAoB,GAAG,CAAC,cAAsB,EAAE,EAAE,CAAC,OAAO,CAAC,EAAE,cAAc,EAAE,CAAC,CAAC;gBACvF,CAAC;gBACD,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,OAAO,EAAE,QAAQ,CAAC,cAAc,CAAC,CAAC;YAC5E,CAAC,CAAC,CAAC;QACL,CAAC;KACF,CAAC;AACJ,CAAC"}
+{"version":3,"file":"execution-context.js","sourceRoot":"","sources":["../../src/http/execution-context.ts"],"names":[],"mappings":"AAAA,gDAAgD;AAChD,OAAO,EAAE,cAAc,EAAE,sBAAsB,EAAE,MAAM,mCAAmC,CAAC;AAK3F;;;;;;;;;GASG;AACH,MAAM,UAAU,qBAAqB,CACnC,IAAiB,EACjB,EAAkB,EAClB,OAAe;IAEf,MAAM,eAAe,GAAG,CAAC,IAAuB,EAAE,EAAE;QAClD,MAAM,gBAAgB,GAAG,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC;QACjD,MAAM,KAAK,GAAG,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,gBAAgB,CAAC,CAAC;QACvD,IAAI,CAAC,KAAK;YAAE,OAAO;QACnB,KAAK,CAAC,eAAe,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QACnC,6DAA6D;QAC7D,mEAAmE;QACnE,6DAA6D;QAC7D,KAAK,CAAC,OAAO,GAAG,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,QAAQ,EAAE,IAAI,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,UAAU,CAAC,EAAE,CAAC,CAAC;QAChF,kEAAkE;QAClE,kEAAkE;QAClE,mEAAmE;QACnE,iDAAiD;QACjD,MAAM,UAAU,GAAG,KAAK,CAAC,UAAU,IAAI,CAAC,CAAC;QACzC,MAAM,QAAQ,GAAG,UAAU,IAAI,CAAC;YAC9B,CAAC,CAAC,IAAI,CAAC,QAAQ;YACf,CAAC,CAAC,sBAAsB,CAAC;gBACrB,UAAU;gBACV,YAAY,EAAE,KAAK,CAAC,YAAY,IAAI,CAAC;gBACrC,cAAc,EAAE,KAAK,CAAC,cAAc,IAAI,CAAC;gBACzC,SAAS,EAAE,KAAK,CAAC,SAAS;gBAC1B,KAAK,EAAE,IAAI,CAAC,GAAG,EAAE;gBACjB,eAAe,EAAE,KAAK,CAAC,eAAe;gBACtC,OAAO,EAAE,KAAK,CAAC,OAAO;aACvB,CAAC,CAAC;QACP,IAAI,CAAC,aAAa,CAAC,qBAAqB,CAAC,gBAAgB,EAAE,QAAQ,CAAC,CAAC;QACrE,IAAI,CAAC,MAAM,CAAC,aAAa,CAAC;YACxB,OAAO,EAAE,gBAAgB;YACzB,SAAS,EAAE,CAAC;YACZ,SAAS,EAAE,IAAI,CAAC,SAAS;YACzB,KAAK,EAAE,IAAI,CAAC,KAAK;SAClB,CAAC,CAAC;QACH,IAAI,IAAI,CAAC,WAAW,KAAK,SAAS,EAAE,CAAC;YACnC,IAAI,CAAC,MAAM,CAAC,eAAe,CAAC;gBAC1B,OAAO,EAAE,gBAAgB;gBACzB,SAAS,EAAE,CAAC;gBACZ,SAAS,EAAE,IAAI,CAAC,WAAW,CAAC,IAAI;gBAChC,OAAO,EAAE,IAAI,CAAC,WAAW,CAAC,EAAE;aAC7B,CAAC,CAAC;QACL,CAAC;IACH,CAAC,CAAC;IAEF,OAAO;QACL,cAAc,EAAE,EAAE;QAClB,MAAM,EAAE,IAAI,CAAC,MAAM;QACnB,MAAM,EAAE,IAAI,CAAC,MAAM;QACnB,iBAAiB,EAAE,EAAE,CAAC,aAAa;QACnC,eAAe,EAAE,CAAC,OAAe,EAAE,EAAE,CAAC,cAAc,CAAC,OAAiC,EAAE,IAAI,CAAC,MAAM,CAAC;QACpG,WAAW,EAAE,OAAO,CAAC,GAAG,CAAC,mBAAmB,CAAC;QAC7C,UAAU,EAAE,SAAS;QACrB,OAAO;QACP,eAAe;QACf,kBAAkB,EAAE,KAAK,EAAE,QAA+B,EAAE,EAAE;YAC5D,OAAO,IAAI,OAAO,CAA6B,CAAC,OAAO,EAAE,EAAE;gBACzD,MAAM,KAAK,GAAG,IAAI,CAAC,aAAa,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;gBAC9C,IAAI,KAAK,EAAE,CAAC;oBACV,0EAA0E;oBAC1E,6CAA6C;oBAC7C,KAAK,CAAC,oBAAoB,GAAG,CAAC,cAAsB,EAAE,EAAE,CAAC,OAAO,CAAC,EAAE,cAAc,EAAE,CAAC,CAAC;gBACvF,CAAC;gBACD,IAAI,CAAC,aAAa,CAAC,oBAAoB,CAAC,OAAO,EAAE,QAAQ,CAAC,cAAc,CAAC,CAAC;YAC5E,CAAC,CAAC,CAAC;QACL,CAAC;KACF,CAAC;AACJ,CAAC"}

package/dist/skills/_shared/polling.md

@@ -64,4 +64,26 @@ Start at 1 s, double each iteration, cap at 30 s. The 1800-second client-side
 timeout is a safety cap; most batches complete in under 60 s. Discover `$PORT`
 at runtime with `mmagent info --json | jq -r .port` (default: 7337).
 
+### Caller-side tool-timeout note
+
+The poll helper's internal `TIMEOUT_S` default is 1800s (30 minutes). If your
+agent's shell tool (e.g. Claude Code's Bash) caps command wall-clock at
+10 minutes by default, the helper will be killed at 10m regardless of
+`TIMEOUT_S` — long-running delegations then appear to "fail" before terminal.
+
+When invoking this poll loop, pick one:
+
+- **Preferred — pass a 30-minute tool timeout explicitly** (e.g. Claude Code
+  Bash accepts `timeout: 1800000`, up to 600000ms/10 min by default; pass the
+  max the tool allows, or bump the tool's allowed ceiling via harness
+  settings).
+- **Alternative — cap the helper to match the tool's limit** by exporting
+  `MMAGENT_POLL_TIMEOUT_S=600` before running the loop. The helper will then
+  exit 124 cleanly at 10 minutes and the caller can decide whether to
+  re-poll or surface the timeout.
+
+Never let the helper run longer than the caller's tool cap — the process
+gets killed mid-poll, the caller sees a generic failure, and diagnostics
+from the `TIMEOUT_S` exit path are lost.
+
 Windows/PowerShell equivalent is planned for a later release.

package/dist/skills/mma-audit/SKILL.md

@@ -10,7 +10,7 @@ when_to_use: >-
   (superpowers:dispatching-parallel-agents, /security-review) points at an audit
   task. Delegate via mmagent so the audit runs on independent workers — your
   main context stays free to synthesize findings.
-version: 3.1.5
+version: 3.1.6
 ---
 
 ## mma-audit

package/dist/skills/mma-clarifications/SKILL.md

@@ -9,7 +9,7 @@ when_to_use: >-
   terminal envelope has `proposedInterpretation` as a string (not a
   NotApplicable sentinel). Read the proposal and call this skill to accept or
   correct it. The batch resumes after the POST returns.
-version: 3.1.5
+version: 3.1.6
 ---
 
 ## mma-clarifications

package/dist/skills/mma-context-blocks/SKILL.md

@@ -10,7 +10,7 @@ when_to_use: >-
   field on mma-delegate / mma-execute-plan / mma-audit / mma-review / mma-verify
   / mma-debug. Cheaper and faster than inlining the same content in every
   request body.
-version: 3.1.5
+version: 3.1.6
 ---
 
 ## mma-context-blocks

package/dist/skills/mma-debug/SKILL.md

@@ -10,7 +10,7 @@ when_to_use: >-
   (superpowers:systematic-debugging) points at the investigation step. Delegate
   the read/reproduce/trace work to a mmagent worker so your main context stays
   focused on the hypothesis and the fix.
-version: 3.1.5
+version: 3.1.6
 ---
 
 ## mma-debug

package/dist/skills/mma-delegate/SKILL.md

@@ -11,7 +11,7 @@ when_to_use: >-
   parallel-safe, and keep main context free. If a plan file exists, use
   mma-execute-plan; if the task is an audit/review/verify/debug, prefer the
   matching mma-* skill instead.
-version: 3.1.5
+version: 3.1.6
 ---
 
 ## mma-delegate
@@ -46,13 +46,11 @@ context block references.
 |---|---|---|---|
 | `tasks` | array | yes | At least one task |
 | `tasks[].prompt` | string | yes | The task instruction |
-| `tasks[].agentType` | string | no | `standard` (default) or `complex` |
+| `tasks[].agentType` | `"standard"` / `"complex"` | no | Worker tier. Default `"standard"` (cheap). Pick `"complex"` when the task is ambiguous, touches many files, is security-sensitive, or a prior standard run came back with `filesWritten: 0` / ran out of turns. Complex workers cost more but finish bigger jobs. |
 | `tasks[].filePaths` | string[] | no | Files the sub-agent focuses on |
 | `tasks[].done` | string | no | Acceptance criteria |
 | `tasks[].contextBlockIds` | string[] | no | IDs from `mma-context-blocks` |
 
-Use `agentType: "complex"` for ambiguous scope or security-sensitive tasks.
-
 ### Full example
 
 ```bash

package/dist/skills/mma-execute-plan/SKILL.md

@@ -11,7 +11,7 @@ when_to_use: >-
   or superpowers:subagent-driven-development / superpowers:executing-plans when
   mmagent is running — delegated workers are cheaper and don't pollute main
   context. Task descriptors must match the plan headings verbatim.
-version: 3.1.5
+version: 3.1.6
 ---
 
 ## mma-execute-plan
@@ -39,7 +39,8 @@ parallel and duplicate descriptors are rejected.
     "/project/docs/plan.md",
     "/project/src/auth/login.ts"
   ],
-  "contextBlockIds": []
+  "contextBlockIds": [],
+  "agentType": "standard"
 }
 ```
 
@@ -49,6 +50,7 @@ parallel and duplicate descriptors are rejected.
 | `context` | string | no | Short additional context not in the plan |
 | `filePaths` | string[] | no | Plan file + relevant source files |
 | `contextBlockIds` | string[] | no | IDs from `mma-context-blocks` |
+| `agentType` | `"standard"` / `"complex"` | no | Worker tier. Default `"standard"` (cheap). Switch to `"complex"` for tasks too large for a standard-tier model to finish in the turn budget (reads many files, produces many edits, or the last run came back with `filesWritten: 0`). |
 
 If the batch reaches `awaiting_clarification`, use `mma-clarifications`
 to confirm or correct the proposed interpretation.

package/dist/skills/mma-retry/SKILL.md

@@ -9,7 +9,7 @@ when_to_use: >-
   you want to re-try the failed indices only. Prefer this over redispatching the
   whole batch or inline-retrying — it's idempotent and keeps the original
   batch's diagnostics intact.
-version: 3.1.5
+version: 3.1.6
 ---
 
 ## mma-retry

package/dist/skills/mma-review/SKILL.md

@@ -9,7 +9,7 @@ when_to_use: >-
   more files OR a methodology skill (superpowers:requesting-code-review,
   /review, /security-review) points at a review task. Delegate the reviewer pass
   to mmagent workers — your main context stays free to decide what to merge.
-version: 3.1.5
+version: 3.1.6
 ---
 
 ## mma-review

package/dist/skills/mma-verify/SKILL.md

@@ -8,7 +8,7 @@ when_to_use: >-
   superpowers:verification-before-completion) wants acceptance-criteria checked
   against implemented work. Delegate the evidence-gathering to mmagent workers —
   each checklist item is verified independently and in parallel.
-version: 3.1.5
+version: 3.1.6
 ---
 
 ## mma-verify

package/dist/skills/multi-model-agent/SKILL.md

@@ -11,7 +11,7 @@ when_to_use: >-
   mmagent is running. Read this once, pick the matching mma-* skill, and
   delegate there. Applies equally whether the user invoked a superpowers
   methodology skill or just asked directly.
-version: 3.1.5
+version: 3.1.6
 ---
 
 ## multi-model-agent overview
@@ -61,6 +61,16 @@ Every request requires `Authorization: Bearer <token>`.
 | `mma-context-blocks` | Register large reused documents to reference by ID |
 | `mma-clarifications` | Confirm or correct the service's proposed interpretation |
 
+### Worker tier: `agentType`
+
+`mma-delegate` and `mma-execute-plan` accept `agentType: "standard" | "complex"`. Default is `"standard"` (cheaper, faster). Pick `"complex"` when:
+
+- The task touches many files or requires multi-step reasoning a smaller model cannot hold in context.
+- A prior standard run came back with `filesWritten: 0` or exhausted its turn budget (visible in the verbose stream or the final envelope's `batchTimings` / `results`).
+- The task is security-sensitive or ambiguous enough that being wrong is costly.
+
+`mma-audit`, `mma-review`, `mma-debug` already default to complex; `mma-verify` already defaults to standard — these are not configurable from the caller and do not need an `agentType` field.
+
 ### General flow
 
 1. Call the appropriate `mma-*` skill → receive `{ batchId }`.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@zhixuan92/multi-model-agent",
-  "version": "3.1.5",
+  "version": "3.1.6",
   "type": "module",
   "license": "MIT",
   "description": "Standalone HTTP server for multi-model-agent. Routes tool-invocation work to Claude, Codex, or OpenAI-compatible sub-agents with async-polling REST dispatch and installable skills for Claude Code, Gemini CLI, Codex CLI, and Cursor.",
@@ -52,7 +52,7 @@
   },
   "dependencies": {
     "@asteasolutions/zod-to-openapi": "^8.5.0",
-    "@zhixuan92/multi-model-agent-core": "^3.1.5",
+    "@zhixuan92/multi-model-agent-core": "^3.1.6",
     "gray-matter": "^4.0.3",
     "minimist": "^1.2.8",
     "zod": "^4.0.0"