musubix3 0.1.1 → 0.1.2

package/.github/plugin/marketplace.json

@@ -3,13 +3,13 @@
   "owner": { "name": "nahisaho" },
   "metadata": {
     "description": "GitHub Copilot CLI specification-driven development skills",
-    "version": "0.1.1"
+    "version": "0.1.2"
   },
   "plugins": [
     {
       "name": "musubix3",
       "source": ".",
-      "version": "0.1.1",
+      "version": "0.1.2",
       "description": "Evidence-driven SDD without duplicating native Copilot capabilities."
     }
   ]

package/.github/skills/sdd-implementation/SKILL.md

@@ -46,11 +46,15 @@ After the work, run `npx musubix3 workflow-record sdd-implementation complete
    Put annotations in the authoritative source and test files, including
    supported non-JS/TS files; never add a proxy file just to increase coverage.
    An annotation establishes a link, not proof that code or tests are correct.
-5. Run the project's actual focused tests/typecheck/build; fix failures.
+5. Before recording Red, run every configured `tdd.redPreflightCommands`
+   formatter/check and let musubix3 enforce that preflight. Then run the
+   project's actual focused tests/typecheck/build; fix failures.
    When mutation evidence is required, use the project's existing lightweight
    mutation mechanism to emit schema-v1 deterministic identities, requirement/
    test linkage, source/test SHA-256, operator/location, and killed status.
-   Never fabricate results or add a large mutation dependency.
+   Never fabricate results or add a large mutation dependency. Use
+   `npx musubix3 mutation doctor --json` to inspect locally available engines
+   and configuration recommendations.
 6. Regenerate `npx musubix3 trace build`, check `trace check --strict`, then
    `npx musubix3 gate --changed`. Configure real command/argument arrays first.
 7. Use Copilot's native review and security-review capabilities when appropriate.

package/.github/skills/sdd-quality/SKILL.md

@@ -8,7 +8,6 @@ Follow the user's input language (日本語 / English). Use native Copilot revie
 security review for their specialist reasoning; this skill does not replace them.
 After the work, run `npx musubix3 workflow-record sdd-quality complete --status
 completed` exactly once.
-
 1. Inspect `.musubix/config.json` and `.musubix/policy-baseline.json` before
    running commands. Only execute trusted project configuration. Commands are
    executable/argument arrays, not shell text. Never edit the baseline without
@@ -18,17 +17,19 @@ completed` exactly once.
    custom report or a built-in Vitest/Jest, pytest, Go test, Cargo or JUnit
    adapter. Require executable native adapter contracts in CI; JUnit targets use
    an exact `@Tag("TEST-*")`, and pytest requires `pytest-json-report`. Review
-   `requiredChecks`, coverage thresholds and architecture rules. Use strict
+   `requiredChecks`, `qualityProfile`, coverage thresholds and architecture rules.
+   Use `recommended` or `release` only when all profile requirements are
+   intentionally configured; profile validation never invents missing evidence. Use strict
    `codeGraph.mode` when unresolved computed module loading must block release,
    and ensure the trusted baseline prevents downgrading it. Configure a
    fresh structured `testReport` for `test-identities`. Do not weaken policy.
 3. Run `npx musubix3 gate --json` or `npx musubix3 gate --changed --json`.
-   `npx musubix3 evidence refresh --json` is the explicit evidence-regeneration
-   form of the same fail-closed gate pipeline. If `input-stability` fails,
+   `evidence refresh --json` runs the same fail-closed pipeline. If `input-stability` fails,
    inspect its per-path added/modified/deleted diagnostics and stop generators
    or formatters before rerunning. Standard dependency/build directories,
-   including Cargo and Maven `target/`, are excluded; source-like generated
-   inputs are not silently ignored.
+   including manifest-scoped Cargo and Maven `target/` and conventional `.venv`
+   or `venv` roots with a regular `pyvenv.cfg`, are excluded; arbitrary source
+   directories and source-like generated inputs are not silently ignored.
    Changed mode reports Git changes and dependent files but conservatively runs
    all checks, including commands. It never treats unrun checks as successful.
    When `tdd` is required, confirm every command has test-scoped `tddArgs`, every
@@ -40,8 +41,9 @@ completed` exactly once.
    Treat workflow records as declarations until `workflow-verify` binds each
    completed declaration to one distinct completed Copilot Skill tool call.
    In strict mode require one final successful result, matching session UUID,
-   causal timestamps, bounded freshness, complete tool lifecycles and canonical
-   transcript hash. Ensure the baseline protects strict/session/freshness policy.
+   causal transcript order, policy-bounded concurrent-event clock skew, bounded
+   freshness, complete tool lifecycles and canonical transcript hash. Ensure the
+   baseline protects strict/session/freshness/event-skew policy.
    When `change-history` is required,
    confirm ordered checkpoints, chained TDD records, requirement-scoped Code
    Graph changes, exact CHANGE requirement enumeration, measurable Acceptance,
@@ -59,7 +61,8 @@ completed` exactly once.
    identities bound to must functional requirements, current source/test
    fingerprints, operator/location, authoritative tests, and killed results.
    Reject survived, skipped, duplicate, conflicting, stale, or unlinked mutants.
-   musubix3 validates evidence; it does not bundle a mutation engine.
+   Run `mutation doctor` for language-aware local probes and remediation;
+   musubix3 validates evidence and does not bundle a mutation engine.
    Missing required tools, commands, artifacts or evidence block readiness.
 5. Use native review/security-review as needed, recording their findings
    separately from machine evidence. Never fabricate review or test results.
@@ -72,4 +75,4 @@ completed` exactly once.
    Ensure the baseline protects CI-required mode, strict OIDC and key binding.
    Never store a private key or overstate OIDC as proof of arbitrary runner work.
 7. Run `npx musubix3 status --json`; stale evidence must be refreshed. Report
-   limitations and residual risks before handoff. No resident watcher or REPL.
+   limitations and residual risks before handoff; no resident watcher or REPL.

package/CHANGELOG.md

@@ -2,6 +2,21 @@
 
 ## Unreleased
 
+## 0.1.2 - 2026-09-07
+
+- Exclude conventional `.venv` and `venv` Python environments identified by a
+  regular `pyvenv.cfg` file from project snapshots and Code Graph indexing,
+  without allowing the marker to hide arbitrary source directories.
+- Add configurable `tdd.redPreflightCommands` so formatters and other plain
+  preflight commands complete before a Red test fingerprint is captured.
+- Add explicit `custom`, `minimal`, `recommended`, and `release` quality
+  profiles with fail-closed requirements for stronger profiles.
+- Accept bounded timestamp skew for causally ordered concurrent Copilot tool
+  events in strict workflow verification, with a policy-protected
+  `workflow.maxEventSkewMs` limit.
+- Add `mutation doctor` with language-aware local engine probes, attempted
+  commands, and actionable configuration recommendations.
+
 ## 0.1.1 - 2026-09-07
 
 - Preserve fail-closed input-stability checks while reporting every added,

package/README-ja.md

@@ -1,6 +1,6 @@
 # musubix3
 
-**v0.1.1 · GitHub Copilot CLI 専用 · Node.js ≥20 · TypeScript · MIT**
+**最新公開版 v0.1.2 · GitHub Copilot CLI 専用 · Node.js ≥20 · TypeScript · MIT**
 
 [English](README.md)
 
@@ -19,16 +19,16 @@ ID の接続や SAT 判定だけで、実装の正しさを保証するもので
 対象プロジェクトで、公開済みパッケージを実行します。
 
 ```sh
-npx musubix3@0.1.1 --version
-npx musubix3@0.1.1 init --dry-run
-npx musubix3@0.1.1 init
+npx musubix3@0.1.2 --version
+npx musubix3@0.1.2 init --dry-run
+npx musubix3@0.1.2 init
 copilot
 ```
 
 バージョンを固定してプロジェクトへインストールする場合は、次を実行します。
 
 ```sh
-npm install --save-dev --save-exact musubix3@0.1.1
+npm install --save-dev --save-exact musubix3@0.1.2
 npx --no-install musubix3 --version
 npx --no-install musubix3 init --dry-run
 npx --no-install musubix3 init
@@ -176,6 +176,7 @@ npx musubix3 status --json
 | `knowledge query <text> [--limit 10]` | TF-IDF/cosine のランキングと陳腐化情報 |
 | `formal generate <file> [--format both\|smt2\|lean]` | SHA-256付きの再現可能なsolver入力生成 |
 | `formal doctor` | Z3、Lean、`lake env lean` の存在とバージョン確認 |
+| `mutation doctor` | 言語別のローカルmutation engine確認と導入推奨 |
 | `formal check <file> [--solver auto\|none\|z3\|lean]` | 明示的なBoolean・条件・数値・時間・状態遷移モデルを検査 |
 | `model-correspondence validate` | Formal JSON→生成trace→正本passing testの証拠を再検証 |
 | `evidence refresh [--changed]` | 同じfail-closed gate pipelineで派生証拠を再生成 |
@@ -327,6 +328,7 @@ version: 1.0.0
 {
   "schemaVersion": 1,
   "language": "auto",
+  "qualityProfile": "custom",
   "commands": [
     { "name": "typecheck", "command": "npm", "args": ["run", "typecheck"], "required": true, "timeoutMs": 120000 },
     {
@@ -342,10 +344,12 @@ version: 1.0.0
   "thresholds": { "design": 1, "implementation": 1, "tests": 1 },
   "formal": { "solver": "none", "minModeledFraction": 0, "timeoutMs": 12000 },
   "mutation": { "mode": "compatible" },
+  "tdd": { "redPreflightCommands": [] },
   "workflow": {
     "mode": "compatible",
     "maxAgeSeconds": 3600,
-    "maxFutureSkewSeconds": 60
+    "maxFutureSkewSeconds": 60,
+    "maxEventSkewMs": 1000
   },
   "attestation": {
     "mode": "local",
@@ -366,6 +370,15 @@ version: 1.0.0
 
 未知の設定キー、不正な閾値、重複コマンドはエラーです。
 glob は `*` / `**` / `?` に対応し、外部依存は `npm:` 接頭辞で表現します。
+`qualityProfile`の既定値は`custom`です。`minimal`はSDDの基本gate、
+`recommended`はstrict Code Graph、TDD、構造化test identityも要求し、
+`release`はformal、mutation、workflow、変更、performance、CI attestationを
+含む完全なrelease checkを要求します。不足設定や弱い設定を証拠で補ったことにはしません。
+`tdd.redPreflightCommands`にはformatter等のplain command名を指定でき、
+Redのtest fingerprintを取得する前に成功が必須です。
+通常ファイルの`pyvenv.cfg`を含む`.venv`と`venv`はsnapshotと
+Code Graphから除外されますが、任意のsource directoryにmarkerを置いても
+除外されません。
 `codeGraph.mode` の既定値は `compatible` で、未解決の計算された
 `import()` / `require()` は警告です。`strict` にするとグラフゲートを阻止する
 エラーになります。信頼済みbaselineが `strict` の場合、`compatible` への
@@ -386,7 +399,9 @@ command reportでpassしたことを要求します。欠落・改変・stale・
 
 `.musubix/policy-baseline.json` は必須チェック、閾値、アーキテクチャ、
 Formalポリシー、mutation mode、workflow strict/session/freshness、CI必須attestation、
-strict OIDC identity/key binding、必須コマンド名の最低条件です。弱体化は拒否され、
+strict OIDC identity/key binding、必須コマンド名の最低条件です。Red preflightを
+要求するbaselineは正規化済み`commands`定義も保持し、同名formatterへの
+すり替えを拒否します。弱体化は拒否され、
 変更ゲート中のbaseline変更には独立承認が必要です。CODEOWNERS等で保護してください。
 
 **信頼した設定だけを実行してください。** ゲートは環境変数を継承し、シェルを

package/README.md

@@ -1,6 +1,6 @@
 # musubix3
 
-**v0.1.1 · GitHub Copilot CLI only · Node.js ≥20 · TypeScript · MIT**
+**Latest published v0.1.2 · GitHub Copilot CLI only · Node.js ≥20 · TypeScript · MIT**
 
 [日本語](README-ja.md)
 
@@ -21,16 +21,16 @@ or requirements are satisfiable.
 Run the published package from the target project:
 
 ```sh
-npx musubix3@0.1.1 --version
-npx musubix3@0.1.1 init --dry-run
-npx musubix3@0.1.1 init
+npx musubix3@0.1.2 --version
+npx musubix3@0.1.2 init --dry-run
+npx musubix3@0.1.2 init
 copilot
 ```
 
 For a reproducible project-local installation:
 
 ```sh
-npm install --save-dev --save-exact musubix3@0.1.1
+npm install --save-dev --save-exact musubix3@0.1.2
 npx --no-install musubix3 --version
 npx --no-install musubix3 init --dry-run
 npx --no-install musubix3 init
@@ -184,6 +184,7 @@ validation/gate or requested solver failure, **2** usage, I/O or malformed confi
 | `knowledge query <text> [--limit 10]` | Deterministic TF-IDF/cosine results and staleness flag |
 | `formal generate <file> [--format both\|smt2\|lean]` | Reproducible solver inputs with SHA-256 evidence |
 | `formal doctor` | Probe Z3, Lean, and `lake env lean` availability and versions |
+| `mutation doctor` | Probe language-aware local mutation engines and show setup recommendations |
 | `formal check <file> [--solver auto\|none\|z3\|lean]` | Check the explicit Boolean/conditional/numeric/temporal/transition model |
 | `model-correspondence validate` | Revalidate Formal JSON → generated trace → authoritative passing test evidence |
 | `evidence refresh [--changed]` | Regenerate derived evidence through the same fail-closed gate pipeline |
@@ -328,6 +329,7 @@ Example `.musubix/config.json` (adapt command arguments to your own project):
 {
   "schemaVersion": 1,
   "language": "auto",
+  "qualityProfile": "custom",
   "commands": [
     { "name": "typecheck", "command": "npm", "args": ["run", "typecheck"], "required": true, "timeoutMs": 120000 },
     {
@@ -343,10 +345,12 @@ Example `.musubix/config.json` (adapt command arguments to your own project):
   "thresholds": { "design": 1, "implementation": 1, "tests": 1 },
   "formal": { "solver": "none", "minModeledFraction": 0, "timeoutMs": 12000 },
   "mutation": { "mode": "compatible" },
+  "tdd": { "redPreflightCommands": [] },
   "workflow": {
     "mode": "compatible",
     "maxAgeSeconds": 3600,
-    "maxFutureSkewSeconds": 60
+    "maxFutureSkewSeconds": 60,
+    "maxEventSkewMs": 1000
   },
   "attestation": {
     "mode": "local",
@@ -367,6 +371,16 @@ Example `.musubix/config.json` (adapt command arguments to your own project):
 
 Config is validated strictly; misspelled keys, invalid bounds and duplicate
 commands fail closed. Globs support `*`, `**`, `?`; external imports use `npm:`.
+`qualityProfile` is `custom` by default. `minimal` preserves the core SDD gate,
+`recommended` also requires strict Code Graph, TDD and structured test
+identities, and `release` requires the complete formal, mutation, workflow,
+change, performance and CI-attestation checks. Stronger profiles reject missing
+or weakened settings rather than silently filling in evidence.
+Use `tdd.redPreflightCommands` to reference plain configured formatter commands;
+they must pass before Red captures the authoritative test fingerprint.
+Conventional `.venv` and `venv` Python environments containing a regular
+`pyvenv.cfg` are excluded from source snapshots and Code Graph indexing.
+Arbitrary source directories containing that marker remain tracked.
 `codeGraph.mode` defaults to `compatible`, where unresolved computed
 `import()`/`require()` calls remain warnings. Set it to `strict` to make those
 diagnostics gate-blocking errors. A trusted strict policy baseline prevents
@@ -391,7 +405,9 @@ diagnostic codes are stable English; human status labels include Japanese.
 `.musubix/policy-baseline.json` records minimum required checks, coverage,
 architecture, formal policy, mutation mode, workflow strict/session/freshness settings,
 CI-required attestation and strict OIDC identity/key binding, and required
-command names. Weakening is rejected; changing the baseline in `gate --changed`
+command names. A baseline that requires TDD Red preflights must also include
+their normalized `commands` definitions, which prevents replacing a trusted
+formatter while retaining its name. Weakening is rejected; changing the baseline in `gate --changed`
 requires independent approval. Protect the baseline with review/CODEOWNERS.
 
 **Only run trusted configuration**: gates execute its commands with inherited

package/dist/packages/analysis/src/config.d.ts

@@ -55,6 +55,7 @@ export interface WorkflowConfig {
     expectedSessionId?: string;
     maxAgeSeconds?: number;
     maxFutureSkewSeconds?: number;
+    maxEventSkewMs?: number;
 }
 export interface CodeGraphConfig {
     mode: 'compatible' | 'strict';
@@ -62,9 +63,14 @@ export interface CodeGraphConfig {
 export interface MutationConfig {
     mode: 'compatible' | 'strict';
 }
+export interface TddConfig {
+    redPreflightCommands: string[];
+}
+export type QualityProfile = 'custom' | 'minimal' | 'recommended' | 'release';
 export interface Config {
     schemaVersion: 1;
     language: 'auto' | 'en' | 'ja';
+    qualityProfile: QualityProfile;
     commands: CommandConfig[];
     requiredChecks: string[];
     thresholds: {
@@ -79,17 +85,21 @@ export interface Config {
     codeGraph: CodeGraphConfig;
     formal: FormalConfig;
     mutation: MutationConfig;
+    tdd: TddConfig;
     workflow: WorkflowConfig;
     attestation: AttestationConfig;
 }
 export interface PolicyBaseline {
     schemaVersion: 1;
+    qualityProfile: QualityProfile;
+    commands: CommandConfig[];
     requiredChecks: string[];
     thresholds: Config['thresholds'];
     architecture: Config['architecture'];
     codeGraph: Config['codeGraph'];
     formal: Pick<FormalConfig, 'solver' | 'minModeledFraction'>;
     mutation: MutationConfig;
+    tdd: TddConfig;
     workflow: WorkflowConfig;
     attestation: AttestationConfig;
     requiredCommands: string[];

package/dist/packages/analysis/src/config.js

@@ -4,6 +4,7 @@ export const checkNames = ['requirements', 'design', 'constitution', 'trace', 'g
 export const defaultConfig = {
     schemaVersion: 1,
     language: 'auto',
+    qualityProfile: 'custom',
     commands: [],
     requiredChecks: ['requirements', 'design', 'constitution', 'trace', 'graph', 'commands'],
     thresholds: { design: 1, implementation: 1, tests: 1 },
@@ -11,7 +12,8 @@ export const defaultConfig = {
     codeGraph: { mode: 'compatible' },
     formal: { solver: 'none', minModeledFraction: 0, timeoutMs: 12_000 },
     mutation: { mode: 'compatible' },
-    workflow: { mode: 'compatible', maxAgeSeconds: 3600, maxFutureSkewSeconds: 60 },
+    tdd: { redPreflightCommands: [] },
+    workflow: { mode: 'compatible', maxAgeSeconds: 3600, maxFutureSkewSeconds: 60, maxEventSkewMs: 1000 },
     attestation: {
         mode: 'local',
         maxAgeSeconds: 3600,
@@ -22,12 +24,15 @@ export const defaultConfig = {
 };
 export const defaultPolicyBaseline = {
     schemaVersion: 1,
+    qualityProfile: defaultConfig.qualityProfile,
+    commands: [],
     requiredChecks: [...defaultConfig.requiredChecks],
     thresholds: { ...defaultConfig.thresholds },
     architecture: { forbidCycles: defaultConfig.architecture.forbidCycles, rules: [...defaultConfig.architecture.rules] },
     codeGraph: { ...defaultConfig.codeGraph },
     formal: { solver: defaultConfig.formal.solver, minModeledFraction: defaultConfig.formal.minModeledFraction },
     mutation: { ...defaultConfig.mutation },
+    tdd: { redPreflightCommands: [...defaultConfig.tdd.redPreflightCommands] },
     workflow: { ...defaultConfig.workflow },
     attestation: {
         ...defaultConfig.attestation,
@@ -71,14 +76,51 @@ function validateAdapterArgs(adapter, args, name) {
         throw new Error(`command ${name} args contain adapter-owned argument ${conflict}; remove it because the ${adapter} adapter adds targeted test/report arguments.`);
     }
 }
+const profileRequiredChecks = {
+    minimal: ['requirements', 'design', 'constitution', 'trace', 'graph', 'commands'],
+    recommended: ['requirements', 'design', 'constitution', 'trace', 'graph', 'commands', 'test-identities', 'tdd'],
+    release: [
+        'requirements', 'design', 'constitution', 'trace', 'graph', 'formal', 'model-correspondence',
+        'mutation', 'workflow', 'tdd', 'change-history', 'change-completeness', 'performance',
+        'attestation', 'test-identities', 'commands',
+    ],
+};
+function validateQualityProfile(config) {
+    if (config.qualityProfile === 'custom')
+        return;
+    const missing = profileRequiredChecks[config.qualityProfile]
+        .filter((check) => !config.requiredChecks.includes(check));
+    if (missing.length) {
+        throw new Error(`qualityProfile ${config.qualityProfile} requires checks: ${missing.join(', ')}.`);
+    }
+    if (config.qualityProfile !== 'minimal' && config.codeGraph.mode !== 'strict') {
+        throw new Error(`qualityProfile ${config.qualityProfile} requires codeGraph.mode strict.`);
+    }
+    if (config.qualityProfile === 'release') {
+        if (config.formal.solver === 'none' || config.formal.minModeledFraction <= 0) {
+            throw new Error('qualityProfile release requires a formal solver and positive modeled coverage.');
+        }
+        if (config.mutation.mode !== 'strict')
+            throw new Error('qualityProfile release requires mutation.mode strict.');
+        if (config.workflow.mode !== 'strict')
+            throw new Error('qualityProfile release requires workflow.mode strict.');
+        if (config.attestation.mode !== 'ci-required') {
+            throw new Error('qualityProfile release requires attestation.mode ci-required.');
+        }
+    }
+}
 export function parseConfig(input) {
     const value = object(input, 'config');
-    keys(value, ['schemaVersion', 'language', 'commands', 'requiredChecks', 'thresholds', 'architecture', 'codeGraph', 'formal', 'mutation', 'workflow', 'attestation'], 'config');
+    keys(value, ['schemaVersion', 'language', 'qualityProfile', 'commands', 'requiredChecks', 'thresholds', 'architecture', 'codeGraph', 'formal', 'mutation', 'tdd', 'workflow', 'attestation'], 'config');
     if (value.schemaVersion !== 1)
         throw new Error('Unsupported config schemaVersion; expected 1.');
     const language = optional(value.language, 'auto');
     if (!['auto', 'en', 'ja'].includes(String(language)))
         throw new Error('language must be auto, en, or ja.');
+    const qualityProfile = optional(value.qualityProfile, defaultConfig.qualityProfile);
+    if (!['custom', 'minimal', 'recommended', 'release'].includes(String(qualityProfile))) {
+        throw new Error('qualityProfile must be custom, minimal, recommended, or release.');
+    }
     const requiredChecks = strings(optional(value.requiredChecks, defaultConfig.requiredChecks), 'requiredChecks');
     if (requiredChecks.some((s) => !checkNames.includes(s)))
         throw new Error('Unknown required check.');
@@ -178,6 +220,20 @@ export function parseConfig(input) {
     });
     if (new Set(commands.map((c) => c.name)).size !== commands.length)
         throw new Error('Command names must be unique.');
+    const rawTdd = object(optional(value.tdd, {}), 'tdd');
+    keys(rawTdd, ['redPreflightCommands'], 'tdd');
+    const redPreflightCommands = strings(optional(rawTdd.redPreflightCommands, defaultConfig.tdd.redPreflightCommands), 'tdd.redPreflightCommands');
+    if (new Set(redPreflightCommands).size !== redPreflightCommands.length) {
+        throw new Error('tdd.redPreflightCommands must not contain duplicates.');
+    }
+    for (const name of redPreflightCommands) {
+        const command = commands.find((candidate) => candidate.name === name);
+        if (!command)
+            throw new Error(`tdd.redPreflightCommands references unknown command ${name}.`);
+        if (command.adapter || command.testReport || command.tddReport || command.mutationReport) {
+            throw new Error(`TDD Red preflight command ${name} must be a plain command without test or mutation reports.`);
+        }
+    }
     const architecture = object(optional(value.architecture, {}), 'architecture');
     keys(architecture, ['forbidCycles', 'rules'], 'architecture');
     if (architecture.forbidCycles !== undefined && typeof architecture.forbidCycles !== 'boolean')
@@ -218,7 +274,7 @@ export function parseConfig(input) {
         throw new Error('mutation.mode must be compatible or strict.');
     }
     const rawWorkflow = object(optional(value.workflow, {}), 'workflow');
-    keys(rawWorkflow, ['mode', 'expectedSessionId', 'maxAgeSeconds', 'maxFutureSkewSeconds'], 'workflow');
+    keys(rawWorkflow, ['mode', 'expectedSessionId', 'maxAgeSeconds', 'maxFutureSkewSeconds', 'maxEventSkewMs'], 'workflow');
     const workflowMode = optional(rawWorkflow.mode, defaultConfig.workflow.mode);
     if (!['compatible', 'strict'].includes(String(workflowMode))) {
         throw new Error('workflow.mode must be compatible or strict.');
@@ -241,6 +297,11 @@ export function parseConfig(input) {
         || workflowMaxFutureSkewSeconds < 0 || workflowMaxFutureSkewSeconds > 600) {
         throw new Error('workflow.maxFutureSkewSeconds must be 0..600.');
     }
+    const workflowMaxEventSkewMs = optional(rawWorkflow.maxEventSkewMs, defaultConfig.workflow.maxEventSkewMs);
+    if (typeof workflowMaxEventSkewMs !== 'number' || !Number.isInteger(workflowMaxEventSkewMs)
+        || workflowMaxEventSkewMs < 0 || workflowMaxEventSkewMs > 60_000) {
+        throw new Error('workflow.maxEventSkewMs must be 0..60000.');
+    }
     const rawAttestation = object(optional(value.attestation, {}), 'attestation');
     keys(rawAttestation, [
         'mode', 'repository', 'maxAgeSeconds', 'maxFutureSkewSeconds', 'trustedPublicKeys', 'githubOidc',
@@ -307,17 +368,24 @@ export function parseConfig(input) {
         && rawAttestation.repository !== rawGithubOidc.repository) {
         throw new Error('attestation.githubOidc.repository must match attestation.repository.');
     }
-    return {
-        schemaVersion: 1, language: language, commands, requiredChecks, thresholds,
+    const config = {
+        schemaVersion: 1,
+        language: language,
+        qualityProfile: qualityProfile,
+        commands,
+        requiredChecks,
+        thresholds,
         architecture: { forbidCycles: architecture.forbidCycles !== false, rules },
         codeGraph: { mode: codeGraphMode },
         formal: { solver: solver, minModeledFraction, timeoutMs: formalTimeoutMs },
         mutation: { mode: mutationMode },
+        tdd: { redPreflightCommands },
         workflow: {
             mode: workflowMode,
             ...(typeof rawWorkflow.expectedSessionId === 'string' ? { expectedSessionId: rawWorkflow.expectedSessionId } : {}),
             maxAgeSeconds: workflowMaxAgeSeconds,
             maxFutureSkewSeconds: workflowMaxFutureSkewSeconds,
+            maxEventSkewMs: workflowMaxEventSkewMs,
         },
         attestation: {
             mode: mode,
@@ -338,6 +406,8 @@ export function parseConfig(input) {
                 : { mode: 'off' },
         },
     };
+    validateQualityProfile(config);
+    return config;
 }
 export async function loadConfig(root) {
     const path = '.musubix/config.json';
@@ -348,31 +418,36 @@ export async function loadConfig(root) {
 export function parsePolicyBaseline(input) {
     const value = object(input, 'policy baseline');
     keys(value, [
-        'schemaVersion', 'requiredChecks', 'thresholds', 'architecture', 'codeGraph', 'formal',
-        'mutation', 'workflow', 'attestation', 'requiredCommands',
+        'schemaVersion', 'qualityProfile', 'commands', 'requiredChecks', 'thresholds', 'architecture', 'codeGraph', 'formal',
+        'mutation', 'tdd', 'workflow', 'attestation', 'requiredCommands',
     ], 'policy baseline');
     if (value.schemaVersion !== 1)
         throw new Error('Unsupported policy baseline schemaVersion; expected 1.');
     const parsed = parseConfig({
         schemaVersion: 1,
+        qualityProfile: value.qualityProfile,
         requiredChecks: value.requiredChecks,
         thresholds: value.thresholds,
         architecture: value.architecture,
         codeGraph: value.codeGraph,
         formal: value.formal,
         mutation: value.mutation,
+        tdd: value.tdd,
         workflow: value.workflow,
         attestation: value.attestation,
-        commands: [],
+        commands: value.commands,
     });
     return {
         schemaVersion: 1,
+        qualityProfile: parsed.qualityProfile,
+        commands: parsed.commands,
         requiredChecks: parsed.requiredChecks,
         thresholds: parsed.thresholds,
         architecture: parsed.architecture,
         codeGraph: parsed.codeGraph,
         formal: { solver: parsed.formal.solver, minModeledFraction: parsed.formal.minModeledFraction },
         mutation: parsed.mutation,
+        tdd: parsed.tdd,
         workflow: parsed.workflow,
         attestation: parsed.attestation,
         requiredCommands: strings(optional(value.requiredCommands, []), 'requiredCommands'),
@@ -386,6 +461,10 @@ export async function loadPolicyBaseline(root) {
 }
 export function policyDiagnostics(config, baseline, changed = null) {
     const diagnostics = [];
+    const profileRank = { custom: 0, minimal: 1, recommended: 2, release: 3 };
+    if (profileRank[config.qualityProfile] < profileRank[baseline.qualityProfile]) {
+        diagnostics.push(error('POLICY_QUALITY_PROFILE', `qualityProfile cannot be weakened below ${baseline.qualityProfile}.`));
+    }
     for (const check of baseline.requiredChecks) {
         if (!config.requiredChecks.includes(check))
             diagnostics.push(error('POLICY_REQUIRED_CHECK', `Required check ${check} was removed from the trusted baseline.`));
@@ -413,6 +492,16 @@ export function policyDiagnostics(config, baseline, changed = null) {
     if (baseline.mutation.mode === 'strict' && config.mutation.mode !== 'strict') {
         diagnostics.push(error('POLICY_MUTATION_MODE', 'mutation.mode cannot be weakened below strict as required by the trusted baseline.'));
     }
+    for (const command of baseline.tdd.redPreflightCommands) {
+        const baselineCommand = baseline.commands.find((candidate) => candidate.name === command);
+        const configuredCommand = config.commands.find((candidate) => candidate.name === command);
+        if (!config.tdd.redPreflightCommands.includes(command) || !configuredCommand) {
+            diagnostics.push(error('POLICY_TDD_PREFLIGHT', `TDD Red preflight command ${command} was removed from the trusted baseline.`));
+        }
+        else if (baselineCommand && JSON.stringify(configuredCommand) !== JSON.stringify(baselineCommand)) {
+            diagnostics.push(error('POLICY_TDD_PREFLIGHT', `TDD Red preflight command ${command} differs from the trusted baseline definition.`));
+        }
+    }
     if (baseline.workflow.mode === 'strict') {
         if (config.workflow.mode !== 'strict') {
             diagnostics.push(error('POLICY_WORKFLOW_MODE', 'workflow.mode cannot be weakened below strict as required by the trusted baseline.'));
@@ -427,6 +516,9 @@ export function policyDiagnostics(config, baseline, changed = null) {
         if ((config.workflow.maxFutureSkewSeconds ?? 60) > (baseline.workflow.maxFutureSkewSeconds ?? 60)) {
             diagnostics.push(error('POLICY_WORKFLOW_FUTURE_SKEW', 'workflow.maxFutureSkewSeconds is weaker than the trusted baseline.'));
         }
+        if ((config.workflow.maxEventSkewMs ?? 1000) > (baseline.workflow.maxEventSkewMs ?? 1000)) {
+            diagnostics.push(error('POLICY_WORKFLOW_EVENT_SKEW', 'workflow.maxEventSkewMs is weaker than the trusted baseline.'));
+        }
     }
     if (baseline.attestation.mode === 'ci-required') {
         if (config.attestation.mode !== 'ci-required') {