@zhixuan92/multi-model-agent 4.4.0 → 4.5.1

package/README.md

@@ -88,7 +88,7 @@ Two ways — pick one:
 
 ```bash
 mmagent serve                          # 127.0.0.1:7337 by default
-curl -s http://localhost:7337/health   # → {"ok":true,"version":"4.4.0",...}
+curl -s http://localhost:7337/health   # → {"ok":true,"version":"4.5.1",...}
 ```
 
 For an always-on background install (survives reboots): [launchd / systemd templates](./scripts/README.md).
@@ -287,14 +287,12 @@ Full design rationale: [DIRECTION.md](https://github.com/zhixuan312/multi-model-
 | TLS `handshake_failure` to a known-good telemetry endpoint | Local DNS cache is stale. `sudo dscacheutil -flushcache && sudo killall -HUP mDNSResponder` (macOS); restart the daemon so its Node process re-resolves |
 | Local telemetry queue stops draining | Daemon's flusher is in exponential backoff after a transport failure (capped at 1 hr). Restart the daemon to force an immediate boot-flush |
 
-## What's new in 4.4.0
+## What's new in 4.5.1
 
-- **Session-based provider boundary.** `Provider.openSession() → Session.send() → TurnResult` replaces the 1,559-line `RunnerShell` chain. All providers (claude-agent-sdk, codex CLI, `@openai/agents`) now expose only `openSession`; cost / termination-reason mapping cannot diverge.
-- **Five-stage lifecycle.** Stages collapse to `implementing → review → rework → annotating → committing`. Single `HUMAN_LABEL` source of truth for headline labels.
-- **Codex token accounting fixed (≈4× cost over-report on cached turns).** OpenAI/codex emit `input_tokens` as GROSS including `cached_input_tokens`; Anthropic emits NET. The codex adapter now subtracts the cached subset before pricing.
-- **`X-MMA-Main-Model` required again.** Auto-detect was reading JSONL files written by our own claude-tier workers and returning the worker's model as "main". The calling client is the only reliable source. Returns `400 main_model_required` if missing.
-- **LLM `verify` tool removed.** Verification is `verifyCommand` only (deterministic shell command run after the worker).
-- **Telemetry clamp ceilings raised** for 2026-era usage: per-stage input/cached `5M → 100M`, output `500K → 2M`, per-stage cost `$100 → $500`, per-task cost `$800 → $5000`.
+- **Windows codex spawn fix.** `codex-cli-session` routes through `cross-spawn` so Node can resolve `codex.cmd` / `.bat` / `.ps1` shims on Windows without falling back to `shell: true` (which would mangle the `-c model_providers.X={…}` argument block on `cmd.exe`). Linux/macOS is a no-op passthrough. Fixes `spawn codex ENOENT` on Windows 4.5.0 daemons.
+- **`mma-audit` plan subtype: 9 → 12 perspectives.** New SPEC COVERAGE (reads upstream spec from a registered context block), PLACEHOLDER LANGUAGE, and PLAN SKELETON perspectives. Grouped as EXTERNAL CODEBASE COHERENCE (1–8), INTRA-PLAN STRUCTURE (9, 11, 12), and SPEC ALIGNMENT (10).
+- **`mma-audit` spec subtype: 7 → 9 criteria.** New PLACEHOLDER-SCAN and DESIGN-DECOMPOSITION-PRESENT criteria; existing SCOPE-EXPLICITNESS extended to flag multi-subsystem specs needing decomposition.
+- **Recipe F (Spec-then-plan-then-execute) updated** in `mma-audit` SKILL.md — register the spec via `mma-context-blocks` between writing-plans and the plan-audit so perspective 10 fires. No schema or wire-shape changes.
 
 Full history: [CHANGELOG](https://github.com/zhixuan312/multi-model-agent/blob/master/CHANGELOG.md).
 

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

@@ -12,7 +12,7 @@ when_to_use: >-
   (superpowers:dispatching-parallel-agents, /security-review) points at one AND
   mmagent is running. Audit on PROSE/SPEC docs — use mma-review for source code.
   Audit a CODE-EXECUTION PLAN against the codebase — use subtype=plan.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # mma-audit
@@ -27,7 +27,7 @@ version: 4.4.0
 |---|---|---|
 | A general prose artifact (design doc, recommendation, post-mortem, README) | `subtype: 'default'` | Comprehensive prose-coherence — would a literal-following worker produce the right outcome from this prose alone? Catches ambiguity, contradictions, missing branches, drift, scope-creep. **Does NOT verify against any codebase.** |
 | A **code-execution PLAN** (`docs/superpowers/plans/*.md` or similar) before running it via `mma-execute-plan` | `subtype: 'plan'` | Plan-vs-codebase coherence — for every method / type / file path / signature / import / verify command the plan names, the codebase actually contains it as described. Catches the bug class the prose-coherence audit cannot see (e.g. plan says `registerBlock` but actual interface is `register`). |
-| A **requirement spec** (what we want, why; success criteria) | `subtype: 'spec'` | Requirement-prose executability — every requirement testable, scope explicit, acceptance criteria covered, non-functional requirements captured, decision-trace exposed, conflicts surfaced. |
+| A **requirement spec** (what we want, why; success criteria) | `subtype: 'spec'` | Requirement-prose executability across 9 criteria — testability, scope explicitness AND decomposability, acceptance-criteria coverage, non-functional capture, requirement conflicts, decision-trace, assumption exposure, placeholder scan, and design-decomposition presence (architecture / components / data flow / error handling / testing). |
 | A **SKILL.md** for an `mma-*` skill or comparable agent-facing playbook | `subtype: 'skill'` | Skill-file reader-effectiveness — when-to-use specificity, endpoint contract integrity, example correctness, anti-pattern coverage, link integrity. |
 
 If you want to bias workers toward a narrow lens (security only, performance only, accessibility only), put that in the free-text `background` portion of the prompt — `subtype` is criteria machinery, not a lens selector.
@@ -75,7 +75,7 @@ Either `document` or `filePaths` (or both) must be provided.
 |---|---|
 | `default` (or omit the field) | **General prose — design doc, recommendation, post-mortem, README, brief.** Comprehensive prose-coherence audit. Does NOT verify against any codebase. |
 | `plan` | **Code-execution plans being audited against a real codebase.** Single-file input (the plan markdown). Workers grep / read source files under `cwd` to verify every named symbol / path / signature / import / verify command. Use this BEFORE every `mma-execute-plan` dispatch. |
-| `spec` | **Requirement spec / brainstorming-output / what-we-want prose.** Criteria target testability, scope explicitness, acceptance-criteria coverage, decision-trace, assumption exposure. |
+| `spec` | **Requirement spec / brainstorming-output / what-we-want prose.** 9 criteria target testability, scope explicitness + decomposability, acceptance-criteria coverage, non-functional capture, requirement conflicts, decision-trace, assumption exposure, placeholder scan, and design-decomposition presence. |
 | `skill` | **`SKILL.md` or comparable agent-facing playbook.** Criteria target when-to-use specificity, endpoint contract integrity, example correctness, anti-pattern coverage, link integrity. |
 
 You can run BOTH on a plan: first `spec` or `default` (prose quality), then `plan` (does the plan match the codebase?). They cover orthogonal failure modes.
@@ -88,7 +88,8 @@ When `subtype: 'plan'`:
 
 - `filePaths` MUST contain exactly **one entry** — the plan markdown. Sending zero or 2+ entries → `400 invalid_request` with the message: *"Plan audit takes exactly one filePath (the plan markdown). The worker discovers and verifies source files itself via its tool surface — do not pre-list source files."*
 - `document` (inline content) is not used in plan mode — the plan must be on disk so workers can reference it by `?cwd=`-relative path.
-- The worker runs the sequential-criteria loop with the plan-audit criteria set: PATH EXISTENCE, SYMBOL EXISTENCE, SIGNATURE MATCH, IMPORT GRAPH, TEST HARNESS AVAILABILITY, STEP SEQUENCE WITHIN TASK, CROSS-TASK DEPENDENCIES, VERIFICATION COMMAND VALIDITY.
+- The worker runs the sequential-criteria loop with the plan-audit criteria set across 12 perspectives in three groups: **EXTERNAL CODEBASE COHERENCE** (1 PATH EXISTENCE, 2 SYMBOL EXISTENCE, 3 SIGNATURE MATCH, 4 IMPORT GRAPH, 5 TEST HARNESS AVAILABILITY, 6 STEP SEQUENCE WITHIN TASK, 7 CROSS-TASK DEPENDENCIES, 8 VERIFICATION COMMAND VALIDITY), **INTRA-PLAN STRUCTURE** (9 TASK GRANULARITY, 11 PLACEHOLDER LANGUAGE, 12 PLAN SKELETON), and **SPEC ALIGNMENT** (10 SPEC COVERAGE).
+- To enable perspective 10 (SPEC COVERAGE), register the upstream spec as a context block via `mma-context-blocks` and pass its `blockId` in `contextBlockIds`. Without a spec in context, perspective 10 emits "No findings for this criterion." and the other 11 still run.
 - Read the findings list. Fix the plan and re-audit if any `critical` or `high` plan-audit findings remain.
 
 ## Full example
@@ -192,7 +193,7 @@ This skill is one step in the larger flow described in `multi-model-agent` → "
 
 - **Recipe E — Plan-validate-execute.** Before any `mma-execute-plan` batch, run `mma-audit` with `subtype: 'plan'` on the plan file. Read the findings. If any `critical` / `high` finding survives, fix the plan and re-audit. This catches the bug class where the plan's named methods/files don't actually exist in the codebase — symbols a prose-coherence audit cannot see.
 
-- **Recipe F — Spec-then-plan-then-execute.** When working from a brainstorming spec: `mma-audit` (`subtype: 'spec'`) → fix → `writing-plans` → `mma-audit` (`subtype: 'plan'`) → fix → `mma-execute-plan`. Spec and plan audits catch orthogonal problem classes.
+- **Recipe F — Spec-then-plan-then-execute (the canonical flow).** When working from a brainstorming spec: `mma-audit` (`subtype: 'spec'`) → fix → `writing-plans` → register the spec as a context block via `mma-context-blocks` → `mma-audit` (`subtype: 'plan'`, `contextBlockIds: [specBlockId]`) → fix → `mma-execute-plan`. Spec audit covers requirement-prose executability; plan audit covers BOTH plan-vs-codebase coherence AND plan-vs-spec coverage (perspective 10 fires only when the spec is in context, which is why the context-block step is load-bearing in this recipe).
 
 Anti-pattern alert: **`parallel-rounds-same-target`** (AP1). Three parallel audits on the same document re-flag the same issues without seeing each other's fixes. Run rounds sequentially with a fix between each.
 

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

@@ -12,7 +12,7 @@ when_to_use: >-
   Register once here, then pass the ID via `contextBlockIds` on mma-delegate /
   mma-execute-plan / mma-audit / mma-review / mma-debug / mma-investigate.
   Cheaper and faster than inlining the same content N times.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # mma-context-blocks

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

@@ -10,7 +10,7 @@ when_to_use: >-
   read files, reproduce, trace — OR a methodology skill
   (superpowers:systematic-debugging) points at the investigation step. Delegate
   the read/reproduce/trace; the main agent stays on the hypothesis and the fix.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # mma-debug

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

@@ -11,7 +11,7 @@ when_to_use: >-
   and keep main context free. If a plan file exists → use mma-execute-plan. If
   the task is audit / review / verify / debug / investigate → use the matching
   specialized skill.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # mma-delegate

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

@@ -10,7 +10,7 @@ when_to_use: >-
   superpowers:subagent-driven-development / superpowers:executing-plans —
   workers are cheaper and don't pollute main context. Task descriptors must
   match plan headings verbatim.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # mma-execute-plan

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

@@ -12,7 +12,7 @@ when_to_use: >-
   out mma-investigate (internal) + mma-research (external) in parallel and
   synthesise the results yourself. DO NOT use for convergent single-answer
   questions — those are mma-investigate.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # mma-explore

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

@@ -12,7 +12,7 @@ when_to_use: >-
   git-history queries. OR you are about to read 3+ files / run any grep in main
   context — that's the inline-labor-leakage anti-pattern (AP2); delegate to this
   skill instead.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # mma-investigate

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

@@ -10,7 +10,7 @@ when_to_use: >-
   others do, what published methods exist) AND mmagent is running. Delegate the
   multi-source web/adapter research to a worker so the main context stays on
   judgment. NOT for codebase questions — those are mma-investigate.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # mma-research

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

@@ -10,7 +10,7 @@ when_to_use: >-
   re-try the failed indices only. Prefer this over re-dispatching the whole
   batch or inline-retrying — it's idempotent and preserves the original batch's
   diagnostics.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # mma-retry

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

@@ -10,7 +10,7 @@ when_to_use: >-
   AND mmagent is running. Delegate so each file reviews on its own worker; the
   main agent only decides what to merge. Review on SOURCE CODE — use mma-audit
   for prose specs / configs.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # mma-review

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

@@ -11,7 +11,7 @@ when_to_use: >-
   tasks — AND 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 asked directly.
-version: 4.4.0
+version: 4.5.1
 ---
 
 # multi-model-agent (router)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@zhixuan92/multi-model-agent",
-  "version": "4.4.0",
+  "version": "4.5.1",
   "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.",
@@ -53,7 +53,7 @@
   },
   "dependencies": {
     "@asteasolutions/zod-to-openapi": "^8.5.0",
-    "@zhixuan92/multi-model-agent-core": "^4.4.0",
+    "@zhixuan92/multi-model-agent-core": "^4.5.1",
     "gray-matter": "^4.0.3",
     "minimist": "^1.2.8",
     "proper-lockfile": "^4.1.2",