cclaw-cli 0.51.30 → 1.0.0

package/dist/content/templates.js

@@ -56,26 +56,6 @@ export const ARTIFACT_TEMPLATES = {
 - **Direct path?** (yes/no + one-line justification)
 - **What if we do nothing?** (concrete consequence, not "nothing happens")
 
-## Forcing Questions
-> Minimum 3 questions; each answer MUST contain at least one *specific* token: a concrete name, a role, a number, a repo-relative path, an external link, or a verbatim quote. Vague answers fail the linter.
-
-| # | Forcing question | Specific answer | Decision impact | Q\\<n\\> decision |
-|---|---|---|---|---|
-| 1 |  |  |  | decision: |
-| 2 |  |  |  | decision: |
-| 3 |  |  |  | decision: |
-
-## Premise List
-> ≥2 premises. Each premise must be in the form \`P<n>: <statement> — agreed | disagreed | revised\`. \`revised\` rows must include the revised statement on the next line.
-
-- P1: <statement> — agreed | disagreed | revised
-- P2: <statement> — agreed | disagreed | revised
-
-## Anti-Sycophancy Stamp
-- **Forbidden response openers acknowledged:** yes (no "you're absolutely right", "great point", "absolutely!", etc.)
-- **Posture commitment:** push back with reasoning when premises feel weak; do not perform agreement.
-- **Evidence-that-would-change-the-recommendation:** (one line per premise, or \`- None.\`)
-
 ## How Might We
 - *How might we …?* — one line naming the user, the desired outcome, and the binding constraint.
 
@@ -151,6 +131,11 @@ RECOMMENDATION: <approach letter — one-line rationale, traced to forcing-quest
 - **Approval:** pending
 - **Next-stage handoff:** On standard track, hand this to \`scope\`; on medium track, hand this directly to \`spec\`. Include upstream decisions used, drift, confidence, unresolved questions, risk hints, and non-goals.
 
+## Embedded Grill
+| Question | Recommended answer | Disposition (accept/refine/reject) |
+|---|---|---|
+|  |  |  |
+
 ## Not Doing
 - (3-5 things this brainstorm is *not* committing to — distinct from \`Deferred\`. These will not appear in scope unless the user explicitly opts in.)
 
@@ -175,6 +160,24 @@ ${SEED_SHELF_SECTION}
 - **Assumptions:**
 - **Open questions (or "None"):**
 
+## Victory Detector
+- Loop status file: \`.cclaw/state/early-loop.json\`
+- Iteration: <n>/<maxIterations>
+- Open concerns: <count>
+- Convergence guard: clear | tripped
+- Decision: iterate | ready_for_advance | escalate_for_human_override
+
+## Critic Pass
+After each producer iteration append one JSON line to \`.cclaw/state/early-loop-log.jsonl\`:
+
+${MARKDOWN_CODE_FENCE}json
+{"ts":"<ISO-8601>","runId":"<run-id>","stage":"brainstorm","iteration":1,"concerns":[{"id":"C-1","severity":"important","locator":"Approaches > Row 2","summary":"<specific concern>"}],"resolvedConcernIds":[]}
+${MARKDOWN_CODE_FENCE}
+
+- \`concerns\` lists unresolved concerns after this pass.
+- Keep concern IDs stable across iterations so convergence checks stay meaningful.
+- Include \`resolvedConcernIds\` when a previously open concern is explicitly closed.
+
 ## Learnings
 - None this stage.
 `,
@@ -211,13 +214,6 @@ ${SEED_SHELF_SECTION}
 | Existing-code leverage? |  |  |
 | Reversibility cost? |  |  |
 
-## Dream State Mapping
-- Deep/optional only; omit for compact scope.
-- CURRENT STATE:
-- THIS PLAN:
-- 12-MONTH IDEAL:
-- Alignment verdict:
-
 ## Implementation Alternatives
 | Option | Summary | Effort (S/M/L/XL) | Risk (Low/Med/High) | Pros | Cons | Reuses |
 |---|---|---|---|---|---|---|
@@ -227,27 +223,6 @@ ${SEED_SHELF_SECTION}
 
 RECOMMENDATION: <option letter — one-line rationale tying back to premise challenge and existing-code leverage>
 
-## Failure Modes Registry
-> Universal failure-mode shape — applies to CLI, library, infra, web, batch jobs.
-
-| Codepath | Failure mode | Rescued? (yes/no) | Test? (unit/integration/e2e) | User sees? (message/silent/N/A) | Logged? (level/none) | Q\\<n\\> decision |
-|---|---|---|---|---|---|---|
-|  |  |  |  |  |  | decision: |
-
-## Reversibility Rating
-- Score (1-5, 1 = one-way door / unrecoverable, 5 = trivially reversible):
-- Justification (cite a specific artifact/file or migration step):
-- Rollback plan reference:
-
-## Temporal Interrogation
-- Deep/optional only; omit for compact scope.
-| Time slice | Likely decision pressure | Lock now or defer? | Reason |
-|---|---|---|---|
-| HOUR 1 (foundations) |  |  |  |
-| HOUR 2-3 (core logic) |  |  |  |
-| HOUR 4-5 (integration) |  |  |  |
-| HOUR 6+ (polish/tests) |  |  |  |
-
 ## Scope Contract
 - **Selected mode:** HOLD SCOPE | SELECTIVE EXPANSION | SCOPE EXPANSION | SCOPE REDUCTION
 - **In scope:**
@@ -300,7 +275,7 @@ RECOMMENDATION: <option letter — one-line rationale tying back to premise chal
 |  |  |  |  |
 
 ## Reference Pull
-- Optional evidence from \`/Users/zuevrs/Downloads/references\`; list accepted/rejected ideas or \`Not needed - compact scope\`.
+- Optional evidence from \`<repo-relative references dir>\`; list accepted/rejected ideas or \`Not needed - compact scope\`.
 
 ## Ambitious Alternatives
 - Optional for SCOPE EXPANSION/SELECTIVE; list larger alternatives and disposition.
@@ -381,6 +356,24 @@ ${SEED_SHELF_SECTION}
 - Explicitly excluded:
 - Next-stage handoff: identify whether the next stage is \`design\` (standard track) or \`spec\` (medium track), and list the exact artifacts/decisions it must carry forward.
 
+## Victory Detector
+- Loop status file: \`.cclaw/state/early-loop.json\`
+- Iteration: <n>/<maxIterations>
+- Open concerns: <count>
+- Convergence guard: clear | tripped
+- Decision: iterate | ready_for_advance | escalate_for_human_override
+
+## Critic Pass
+After each producer iteration append one JSON line to \`.cclaw/state/early-loop-log.jsonl\`:
+
+${MARKDOWN_CODE_FENCE}json
+{"ts":"<ISO-8601>","runId":"<run-id>","stage":"scope","iteration":1,"concerns":[{"id":"C-1","severity":"important","locator":"Scope Contract > In Scope","summary":"<specific concern>"}],"resolvedConcernIds":[]}
+${MARKDOWN_CODE_FENCE}
+
+- \`concerns\` lists unresolved concerns after this pass.
+- Keep concern IDs stable across iterations so convergence checks stay meaningful.
+- Include \`resolvedConcernIds\` when a previously open concern is explicitly closed.
+
 ## Learnings
 - None this stage.
 `,
@@ -492,31 +485,17 @@ ${MARKDOWN_CODE_FENCE}
 (failure detection -> rescue action -> user-visible outcome)
 ${MARKDOWN_CODE_FENCE}
 
-## State Machine Diagram
-- Deep add-on; omit for compact design.
+## Deep Diagram Add-on
+- Deep add-on only.
+- type: state-machine | rollback | deployment-sequence | none
+- Choose exactly one marker below when type is not \`none\`.
 
 <!-- diagram: state-machine -->
-
-${MARKDOWN_CODE_FENCE}
-(state transitions for the critical flow lifecycle)
-${MARKDOWN_CODE_FENCE}
-
-## Rollback Flowchart
-- Deep add-on; omit for compact design.
-
 <!-- diagram: rollback-flowchart -->
-
-${MARKDOWN_CODE_FENCE}
-(trigger -> rollback actions -> verification)
-${MARKDOWN_CODE_FENCE}
-
-## Deployment Sequence Diagram
-- Deep add-on; omit for compact design.
-
 <!-- diagram: deployment-sequence -->
 
 ${MARKDOWN_CODE_FENCE}
-(rollout order, guard checks, and verification sequence)
+(deep add-on diagram: state lifecycle OR rollback flow OR deployment sequence)
 ${MARKDOWN_CODE_FENCE}
 
 ## Stale Diagram Audit
@@ -571,32 +550,6 @@ ${MARKDOWN_CODE_FENCE}
 |---|---|---|---|
 |  |  |  | covered/gap |
 
-## ASCII Coverage Diagram
-
-<!-- diagram: ascii-coverage -->
-
-${MARKDOWN_CODE_FENCE}
-entry-point
-  ├── happy path           [★★★]
-  ├── empty input          [★★]
-  ├── error path           [★]
-  ├── concurrency edge     [GAP]
-  ├── slow-network edge    [→E2E]
-  └── perf regression      [→EVAL]
-${MARKDOWN_CODE_FENCE}
-
-> Required marker tokens (at least one each present where applicable): \`[★★★]\` / \`[★★]\` / \`[★]\` / \`[GAP]\` / \`[→E2E]\` / \`[→EVAL]\`. The diagram is the single source of truth for coverage; gaps must be traced into Plan or Spec.
-
-## Regression Iron Rule
-- Iron rule acknowledged: yes — any diff that changes existing behavior gets a regression test added to the plan, no exceptions.
-- Detected behavior changes (or \`- None.\`):
-- Regression test handoff (Plan task ID or \`- None.\`):
-
-## Calibrated Findings
-> Format: \`[P1|P2|P3] (confidence: <n>/10) <repo-relative-path>[:<line>] — <one-line description>\`. Findings with confidence \`< 7\` are suppressed unless severity is \`P1\`.
-
-- (or \`- None this stage.\`)
-
 ## Performance Budget
 | Critical path | Metric | Target | Measurement method |
 |---|---|---|---|
@@ -629,6 +582,9 @@ ${MARKDOWN_CODE_FENCE}
 - Test/performance expectations:
 - Unresolved questions (or \`None\`):
 
+## Long-Term Trajectory
+- (1-3 lines: what comes after this ships, and whether this architecture can absorb that path without major rework. For tactical changes use \`None - tactical change only\`.)
+
 ## Outside Voice Findings
 | ID | Dimension | Finding | Disposition | Rationale |
 |---|---|---|---|---|
@@ -689,8 +645,23 @@ ${SEED_SHELF_SECTION}
 
 **Decisions made:** 0 | **Unresolved:** 0
 
-## Learning Capture Hint
-For meaningful design work, replace the Learnings sentinel with 1-3 JSON learning bullets, for example: \`- {"type":"lesson","trigger":"when design chooses a risky fallback path","action":"record the switch trigger and rollback signal in Spec Handoff","confidence":"medium","domain":"architecture","stage":"design"}\`
+## Victory Detector
+- Loop status file: \`.cclaw/state/early-loop.json\`
+- Iteration: <n>/<maxIterations>
+- Open concerns: <count>
+- Convergence guard: clear | tripped
+- Decision: iterate | ready_for_advance | escalate_for_human_override
+
+## Critic Pass
+After each producer iteration append one JSON line to \`.cclaw/state/early-loop-log.jsonl\`:
+
+${MARKDOWN_CODE_FENCE}json
+{"ts":"<ISO-8601>","runId":"<run-id>","stage":"design","iteration":1,"concerns":[{"id":"C-1","severity":"important","locator":"Engineering Lock > Row 1","summary":"<specific concern>"}],"resolvedConcernIds":[]}
+${MARKDOWN_CODE_FENCE}
+
+- \`concerns\` lists unresolved concerns after this pass.
+- Keep concern IDs stable across iterations so convergence checks stay meaningful.
+- Include \`resolvedConcernIds\` when a previously open concern is explicitly closed.
 
 ## Learnings
 - None this stage.
@@ -773,12 +744,6 @@ For meaningful design work, replace the Learnings sentinel with 1-3 JSON learnin
 |---|---|---|
 |  |  |  |
 
-## Testing Strategy
-- Behaviors covered (not implementation):
-- Integration vs. unit split (and why):
-- Real services vs. doubles (and why):
-- Coverage gaps with rationale (or \`- None.\`):
-
 ## Spec Self-Review
 > Inline pass; fix in place. If a check fails, do not move on without recording the fix.
 
@@ -791,13 +756,6 @@ For meaningful design work, replace the Learnings sentinel with 1-3 JSON learnin
 - Remaining concerns:
   - None
 
-## Reviewer Concerns (convergence guard)
-> Populate ONLY if the spec review loop did not converge after 3 iterations. Each row links a concern to the unresolved review pass.
-
-| ID | Concern | Reviewer / source | Disposition (open/accept/defer) | Rationale |
-|---|---|---|---|---|
-|  |  |  |  |  |
-
 ## Approval
 - Approved by:
 - Date:
@@ -906,23 +864,16 @@ Execution rule: complete and verify each batch before starting the next batch.
   - [ ] Step 4: run test, observe GREEN
   - [ ] Step 5: refactor + commit
 
-## High-Level Technical Design
-> "Directional guidance, not implementation specification." Choose the form that fits the work: pseudo-code grammar, mermaid sequence/state, data-flow ASCII, decision matrix. Skip if the plan is a pure rename/move.
-
-\`\`\`
-(pseudo-code, mermaid, ASCII data flow, or decision matrix)
-\`\`\`
+## Calibrated Findings
+> Recommended for high-risk or multi-batch plans. Use one line per finding, or \`None this stage\`.
+>
+> Canonical format:
+> \`- [P1|P2|P3] (confidence: <n>/10) <path>[:<line>] — <description>\`
 
-## Plan Self-Review
-- [ ] Spec coverage: every spec behavior maps to a unit/task
-- [ ] Placeholder scan (regex on full artifact, not only Task List)
-- [ ] Type/name consistency across units (signatures referenced match definitions)
-- [ ] No silent scope reduction
-- [ ] Confidence per unit recorded (1-10)
-- Patches applied:
-  - None
-- Remaining concerns:
-  - None
+## Regression Iron Rule
+- Iron rule acknowledged: yes
+- Critical regression guardrail:
+- Required pre-merge proof:
 
 ## Execution Handoff
 - **Posture chosen:** Subagent-Driven (recommended) | Inline executor
@@ -930,13 +881,13 @@ Execution rule: complete and verify each batch before starting the next batch.
 - **Subagent recipe (if Subagent-Driven):** \`<harness>\` -> \`<dispatch surface>\` -> \`<agent-definition path>\` (substitute neutral placeholders; full recipes in \`docs/harnesses.md\`)
 - **Inline recipe (if Inline executor):** TDD loop unit-by-unit with batch checkpoints
 
-## No-Placeholder Scan
-- Scanned tokens: \`TODO\`, \`TBD\`, \`FIXME\`, \`<fill-in>\`, \`<your-*-here>\`, \`xxx\`, bare ellipsis in task rows.
-- Hits: 0 (required for WAIT_FOR_CONFIRM to resolve).
-
-## No Scope Reduction Language Scan
-- Scanned phrases: \`v1\`, \`for now\`, \`later\`, \`temporary\`, \`placeholder\`, \`mock for now\`, \`hardcoded for now\`, \`will improve later\`.
-- Hits: 0 (required when Locked Decisions section is non-empty; use LD#hash anchors).
+## Plan Quality Scan
+- Placeholder scan:
+  - Scanned tokens: \`TODO\`, \`TBD\`, \`FIXME\`, \`<fill-in>\`, \`<your-*-here>\`, \`xxx\`, bare ellipsis in task rows.
+  - Hits: 0 (required for WAIT_FOR_CONFIRM to resolve).
+- Scope reduction language scan:
+  - Scanned phrases: \`v1\`, \`for now\`, \`later\`, \`temporary\`, \`placeholder\`, \`mock for now\`, \`hardcoded for now\`, \`will improve later\`.
+  - Hits: 0 (required when Locked Decisions section is non-empty; use LD#hash anchors).
 
 ## WAIT_FOR_CONFIRM
 - Status: pending
@@ -1028,18 +979,27 @@ Execution rule: complete and verify each batch before starting the next batch.
 |---|---|---|---|
 | S-1 |  |  |  |
 
-## Anti-Rationalization Checks
-- [ ] No "test passes immediately" — each new test was watched failing first
-- [ ] No "code before test" reuse from a prior session
-- [ ] No "tests after" backfill instead of RED-first
-- [ ] No "spirit not ritual" overrides
-- Notes (or \`- None this stage.\`):
-
 ## Verification Ladder
 | Slice | Tier reached | Evidence |
 |---|---|---|
 | S-1 |  |  |
 
+## TDD Blocker Taxonomy
+> Fill only when blocked. Choose one canonical class and capture the resume contract.
+>
+> Allowed classes: \`NO_SOURCE_CONTEXT\`, \`NO_TEST_SURFACE\`, \`NO_IMPLEMENTABLE_SLICE\`, \`RED_NOT_EXPRESSIBLE\`, \`NO_VCS_MODE\`.
+
+| Slice | Class | blockedBecause | missingInputs | recommendedRoute | nextCommand | resumeCriteria |
+|---|---|---|---|---|---|---|
+| S-1 |  |  |  |  |  |  |
+
+## Per-Slice Review
+> Only required when \`.cclaw/config.yaml::sliceReview.enabled\` triggers this slice; otherwise record \`not triggered\`.
+
+| Slice | Trigger fired | Spec-Compliance pass | Quality pass | Fulfillment mode (isolated/role-switch/not triggered) |
+|---|---|---|---|---|
+| S-1 |  |  |  |  |
+
 ## Coverage Targets
 | Code type | Target | Current | Command |
 |---|---|---|---|
@@ -1059,11 +1019,78 @@ Execution rule: complete and verify each batch before starting the next batch.
 |---|---|---|---|---|
 | S-1 |  |  |  |  |
 
-## Learning Capture Hint
-For meaningful TDD work, replace the Learnings sentinel with 1-3 JSON learning bullets, for example: \`- {"type":"pattern","trigger":"when a regression only fails after state rewind","action":"keep the RED fixture and add a cycle-log assertion before GREEN","confidence":"medium","domain":"testing","stage":"tdd"}\`
+## Learnings
+- None this stage.
+`,
+    "cohesion-contract.md": `${artifactFrontmatter("tdd")}
+
+# Cohesion Contract — <wave / stage / topic>
+
+## Shared Types & Interfaces
+| Symbol | Path | Signature | Owner slice |
+|---|---|---|---|
+|  |  |  |  |
+
+## Naming Conventions
+- 
+
+## Invariants
+- 
+
+## Integration Touchpoints
+| From slice | To slice | Surface | Integration test name |
+|---|---|---|---|
+|  |  |  |  |
+
+## Behavior Specifications per Slice
+### Slice <n>: <description>
+- test: <name>
+  assert: <one-line assertion>
+  surface: <public interface>
+
+## Status
+| Slice | Implemented | Tests pass | Cohesion verified |
+|---|---|---|---|
+| S-1 | no | no | no |
 
 ## Learnings
 - None this stage.
+`,
+    "cohesion-contract.json": `{
+  "version": 1,
+  "sharedTypes": [
+    {
+      "symbol": "",
+      "path": "",
+      "signature": "",
+      "ownerSlice": ""
+    }
+  ],
+  "touchpoints": [
+    {
+      "fromSlice": "",
+      "toSlice": "",
+      "surface": "",
+      "integrationTestName": ""
+    }
+  ],
+  "slices": [
+    {
+      "sliceId": "S-1",
+      "description": "",
+      "test": "",
+      "assert": "",
+      "surface": "",
+      "implemented": false,
+      "testsPass": false,
+      "cohesionVerified": false
+    }
+  ],
+  "status": {
+    "overall": "pending",
+    "notes": ""
+  }
+}
 `,
     "07-review.md": `${artifactFrontmatter("review")}
 
@@ -1076,15 +1103,13 @@ For meaningful TDD work, replace the Learnings sentinel with 1-3 JSON learning b
 - Open questions:
 - Drift from upstream (or \`None\`):
 
-## Self-Review First
+## Pre-Critic Self-Review
 - [ ] Build/lint/type-check/tests passed locally
 - [ ] Diff matches spec/plan (no scope creep)
 - [ ] Leftover prints / commented code / unused imports removed
 - [ ] Deletion test: each new module justifies its existence
 - Evidence (commands + result):
 - Patches applied (or \`- None.\`):
-
-## Frame the Review Request
 - **Goal:**
 - **Approach:**
 - **Risk areas:**
@@ -1139,6 +1164,12 @@ For meaningful TDD work, replace the Learnings sentinel with 1-3 JSON learning b
 | R-1 | Critical/Important/Suggestion | correctness/security/performance/architecture/external-safety | path:line |  | open/resolved |
 - NO_FINDINGS_ATTESTATION: <required when no findings are reported; cite inspected coverage>
 
+## Lens Coverage
+- Performance: NO_IMPACT | FOUND_<n>
+- Compatibility: NO_IMPACT | FOUND_<n>
+- Observability: NO_IMPACT | FOUND_<n>
+- Security: routed to security-reviewer (always separate)
+
 ## Security Sweep Attestation
 - Result: findings | NO_CHANGE_ATTESTATION | NO_SECURITY_IMPACT
 - Inspected surfaces:
@@ -1160,7 +1191,7 @@ For meaningful TDD work, replace the Learnings sentinel with 1-3 JSON learning b
 
 ## Review Readiness Snapshot
 
-- Victory Detector: pass | fail (Layer 1, Layer 2, security sweep, structured findings, trace evidence, unresolved-critical status)
+- Victory Detector: pass | fail (Layer 1, Layer 2, security sweep, structured findings, acceptance/reproduction coverage evidence, unresolved-critical status)
 - Completed checks: Layer 1, Layer 2 tags, security sweep, schema validation
 - Delegation log: \`.cclaw/state/delegation-log.json\` required/completed/waived/pending
 - Staleness signal: commit at last review pass vs current commit
@@ -1175,11 +1206,10 @@ For meaningful TDD work, replace the Learnings sentinel with 1-3 JSON learning b
 - Adversarial review: not triggered | pass | fail
 - Overall: complete | concerns | blocked
 
-## Trace Matrix Check
-- Command: \`cclaw internal trace-matrix\` when the active track enforces it; otherwise record direct AC/reproduction-slice coverage.
-- Orphaned criteria: 0
-- Orphaned source items: 0 or \`N/A - direct spec/reproduction coverage\`
-- Orphaned tests: 0
+## Coverage Check
+- AC/source-item/slice coverage rationale:
+- Orphaned source items: none | explain gap
+- Orphaned tests: none | explain gap
 - Evidence ref:
 
 ## Verification Command Discovery
@@ -1201,9 +1231,6 @@ For meaningful TDD work, replace the Learnings sentinel with 1-3 JSON learning b
 ## Final Verdict
 - APPROVED | APPROVED_WITH_CONCERNS | BLOCKED
 
-## Learning Capture Hint
-For meaningful review work, replace the Learnings sentinel with 1-3 JSON learning bullets, for example: \`- {"type":"lesson","trigger":"when security sweep finds no issues but touches trust boundaries","action":"record NO_SECURITY_IMPACT with inspected surfaces and rationale","confidence":"medium","domain":"security","stage":"review"}\`
-
 ## Learnings
 - None this stage.
 `,
@@ -1257,6 +1284,12 @@ For meaningful review work, replace the Learnings sentinel with 1-3 JSON learnin
 - Type-check:
 - Working tree clean:
 
+## Architect Cross-Stage Verification
+- Skill: architect-cross-stage-verification
+- Result: CROSS_STAGE_VERIFIED | DRIFT_DETECTED | BLOCKED
+- Evidence refs:
+- Drift summary:
+
 ## Base Branch Determination
 - Command run: \`git merge-base HEAD main || git merge-base HEAD master\`
 - Base branch:
@@ -1438,7 +1471,7 @@ Track-specific skips are allowed only when \`flow-state.track\` + \`skippedStage
 
 - No completion claim without fresh command evidence in this turn.
 - Do not mark gates passed from memory.
-- Keep evidence in \`.cclaw/artifacts/\`; archive via \`npx cclaw-cli archive\`.
+- Keep evidence in \`.cclaw/artifacts/\`; archive through closeout via \`/cc\` or cancel early via \`node .cclaw/hooks/cancel-run.mjs\`.
 
 ## Delegation And Approvals
 

package/dist/content/tree-command.js

@@ -31,7 +31,7 @@ Do not modify state in this command. It is a pure read/render operation.
    \`fulfillmentMode\` / \`evidenceRefs\`.
 3. Inspect \`${artifactsPath()}\` for per-stage artifact presence and stale copies,
    and for the retro artifact \`09-retro.md\`.
-4. Use \`npx cclaw-cli doctor --explain\` for harness capability status when needed.
+4. Use \`npx cclaw-cli sync\` for harness capability status when needed.
 5. Render one compact tree:
    - stage marker: passed/current/pending/skipped/stale,
    - gates summary,
@@ -47,7 +47,7 @@ Do not modify state in this command. It is a pure read/render operation.
      then \`next\`; the transient \`archived\` substate surfaces only if the
      archive step failed mid-run.
 7. Append a \`harnesses:\` branch. For each installed harness derive the tier
-   and fallback from cclaw capability metadata; use \`npx cclaw-cli doctor --explain\`
+   and fallback from cclaw capability metadata; use \`npx cclaw-cli sync\`
    for remediation details when needed.
 8. If rewind records exist in \`${rewindLogPath()}\`, include latest rewind note in footer.
 

package/dist/content/utility-skills.d.ts

@@ -18,13 +18,13 @@ export declare const LANGUAGE_RULE_PACK_FILES: {
 };
 /**
  * Folder (relative to runtime root) that holds every enabled language rule
- * pack. A single folder keeps discovery trivial for hooks and for `doctor`.
+ * pack. A single folder keeps discovery trivial for hooks and for `sync`.
  */
 export declare const LANGUAGE_RULE_PACK_DIR: readonly ["rules", "lang"];
 export declare const LANGUAGE_RULE_PACK_GENERATORS: Record<string, () => string>;
 /**
  * Legacy per-language folders under `.cclaw/skills/` used in v0.7.0. Listed
- * here so `cclaw sync` and `doctor` can surface drift and the installer can
+ * here so `cclaw sync` and `sync` can surface drift and the installer can
  * clean them up after the move to `.cclaw/rules/lang/`.
  */
 export declare const LEGACY_LANGUAGE_RULE_PACK_FOLDERS: readonly ["language-typescript", "language-python", "language-go"];