@hegemonart/get-design-done 1.41.5 → 1.43.0

package/skills/quality-gate/SKILL.md

@@ -20,9 +20,9 @@ writes:
 
 ## Role
 
-You are the Stage 4.5 gate that runs between `/gdd:design` and `/gdd:verify`. You answer one question: *does this project's own quality tooling pass against the current working tree?* You are NOT a design checker, an a11y checker, or a verifier — you are a thin façade over the project's `lint` / `typecheck` / `test` / visual-regression scripts. Verify refuses entry when those scripts fail.
+You are the Stage 4.5 gate that runs between `/gdd:design` and `/gdd:verify`. You answer one question: *does this project's own quality tooling pass against the current working tree?* You are NOT a design checker, an a11y checker, or a verifier - you are a thin façade over the project's `lint` / `typecheck` / `test` / visual-regression scripts. Verify refuses entry when those scripts fail.
 
-You write exactly two artifacts: the `<quality_gate>` block in `.design/STATE.md`, and lifecycle events to `.design/events.jsonl`. You never block on timeout. You never block on a "skipped" result. You only mark `status="fail"` when the fix loop reaches `max_iters` — even then YOU exit successfully (verify is the consumer that refuses entry).
+You write exactly two artifacts: the `<quality_gate>` block in `.design/STATE.md`, and lifecycle events to `.design/events.jsonl`. You never block on timeout. You never block on a "skipped" result. You only mark `status="fail"` when the fix loop reaches `max_iters` - even then YOU exit successfully (verify is the consumer that refuses entry).
 
 ## Configuration
 
@@ -34,57 +34,57 @@ Read once at start from `.design/config.json` (all optional; defaults in parens)
 | `quality_gate.timeout_seconds` | `600` | Total wall-clock budget for Step 2. On timeout: warn + proceed (D-07). |
 | `quality_gate.max_iters` | `3` | Hard cap on Step 4 fix-loop iterations. |
 
-## Step 1 — Detection chain (D-06 3-tier fallback)
+## Step 1 - Detection chain (D-06 3-tier fallback)
 
 Stop at the first tier that produces ≥ 1 command:
 
 1. **Authoritative config.** If `.design/config.json` has `quality_gate.commands` non-empty, use verbatim.
-2. **Auto-detect from `package.json#scripts`** — match against allowlist: `lint`, `typecheck`, `tsc` (only if `typecheck` absent), `test`, `chromatic`, `test:visual`, `lint:design` (Phase 41 — the `gdd-detect` deterministic anti-pattern gate, alongside `axe`/`pa11y`/`lighthouse`). Exclude by name: `test:e2e`, `test:integration` (if separate `test`), anything starting `dev:`, `build:`, `start:`. Run via `npm run <name>` unless `quality_gate.package_manager` overrides.
+2. **Auto-detect from `package.json#scripts`** - match against allowlist: `lint`, `typecheck`, `tsc` (only if `typecheck` absent), `test`, `chromatic`, `test:visual`, `lint:design` (Phase 41 - the `gdd-detect` deterministic anti-pattern gate, alongside `axe`/`pa11y`/`lighthouse`). Exclude by name: `test:e2e`, `test:integration` (if separate `test`), anything starting `dev:`, `build:`, `start:`. Run via `npm run <name>` unless `quality_gate.package_manager` overrides.
 3. **Skip with notice.** Emit `quality_gate_skipped` (Step 6) and write a `<run/>` with `status="skipped"`. Verify treats skipped as non-blocking.
 
-## Step 2 — Parallel run
+## Step 2 - Parallel run
 
-Emit `quality_gate_started`. Spawn each command in a separate `Bash`; collect `{command, exit_code, stdout, stderr}`. Wall-clock budget is `timeout_seconds` (default 600). On timeout: emit `quality_gate_timeout`, mark `status="timeout"`, skip Steps 3–4, proceed to Step 5. Exit successfully — verify treats timeout as a warn.
+Emit `quality_gate_started`. Spawn each command in a separate `Bash`; collect `{command, exit_code, stdout, stderr}`. Wall-clock budget is `timeout_seconds` (default 600). On timeout: emit `quality_gate_timeout`, mark `status="timeout"`, skip Steps 3–4, proceed to Step 5. Exit successfully - verify treats timeout as a warn.
 
-## Step 3 — Classification
+## Step 3 - Classification
 
 Spawn `quality-gate-runner` agent via `Task` with payload `{outputs: [{command, exit_code, stderr}, ...]}`. Agent returns `{status: "pass"|"fail", classified_failures: {lint, type, test, visual}}`. `pass` → Step 5. `fail` → Step 4.
 
-## Step 4 — Fix loop (D-08)
+## Step 4 - Fix loop (D-08)
 
 If `iteration >= max_iters`: emit `quality_gate_fail`, mark `status="fail"`, Step 5, exit successfully. Verify-entry refuses on `fail`; YOU do not throw.
 
-Else: increment `iteration`, emit `quality_gate_iteration`, spawn `design-fixer` via `Task` with classified failures + original outputs. After fixer returns, restart from Step 2 (re-run all commands — fixes can introduce regressions).
+Else: increment `iteration`, emit `quality_gate_iteration`, spawn `design-fixer` via `Task` with classified failures + original outputs. After fixer returns, restart from Step 2 (re-run all commands - fixes can introduce regressions).
 
-## Step 5 — STATE write
+## Step 5 - STATE write
 
-Mutate `state.quality_gate.run` to `{started_at, completed_at, status, iteration, commands_run, extra_attrs:{}}`. Persist via `mcp__gdd_state__set_quality_gate` or `apply()` mutator from `sdk/state/mutator.ts` — identical on-disk shape.
+Mutate `state.quality_gate.run` to `{started_at, completed_at, status, iteration, commands_run, extra_attrs:{}}`. Persist via `mcp__gdd_state__set_quality_gate` or `apply()` mutator from `sdk/state/mutator.ts` - identical on-disk shape.
 
-## Step 6 — Event emission (D-09)
+## Step 6 - Event emission (D-09)
 
-Use `appendEvent` from `sdk/event-stream/index.ts` — persist-first / broadcast-second; never throws on persist path. `ts` / `cycle` / `stage` are stamped by the writer. Six event types (one per lifecycle position):
+Use `appendEvent` from `sdk/event-stream/index.ts` - persist-first / broadcast-second; never throws on persist path. `ts` / `cycle` / `stage` are stamped by the writer. Six event types (one per lifecycle position):
 
 | Event | When | Payload |
 |-------|------|---------|
 | `quality_gate_started` | Step 2 entry, once | `commands`, `timeout_seconds`, `max_iters` |
 | `quality_gate_iteration` | Step 4 retry (iter ≥ 2) | `iteration` |
-| `quality_gate_pass` | Step 3 returned pass — terminal | `iteration`, `commands_run` |
-| `quality_gate_fail` | Step 4 hit `max_iters` — terminal | `iteration`, `classified_failures` |
-| `quality_gate_timeout` | Step 2 budget elapsed — terminal warn | `unfinished_commands` |
-| `quality_gate_skipped` | Step 1 tier 3 — terminal no-op | `reason` |
+| `quality_gate_pass` | Step 3 returned pass - terminal | `iteration`, `commands_run` |
+| `quality_gate_fail` | Step 4 hit `max_iters` - terminal | `iteration`, `classified_failures` |
+| `quality_gate_timeout` | Step 2 budget elapsed - terminal warn | `unfinished_commands` |
+| `quality_gate_skipped` | Step 1 tier 3 - terminal no-op | `reason` |
 
-`appendEvent` swallows persist failures — events are observability, not correctness. STATE.md (Step 5) is the durable record.
+`appendEvent` swallows persist failures - events are observability, not correctness. STATE.md (Step 5) is the durable record.
 
 ## Output
 
-Emit one JSON object on stdout: `{status, iteration, commands_run, started_at, completed_at}`. Shell exit code `0` on every terminal status — `fail` included. Verify-entry is the sole consumer that acts on `fail`.
+Emit one JSON object on stdout: `{status, iteration, commands_run, started_at, completed_at}`. Shell exit code `0` on every terminal status - `fail` included. Verify-entry is the sole consumer that acts on `fail`.
 
 ## Constraints
 
-- Do not prune the command list across iterations — re-run everything in Step 2.
+- Do not prune the command list across iterations - re-run everything in Step 2.
 - Do not spawn `quality-gate-runner` more than once per iteration.
 - Do not read/write any STATE block other than `<quality_gate>` (and `<position>.last_checkpoint`).
-- Do not invoke verify or design — Stage 4.5 sits strictly between.
+- Do not invoke verify or design - Stage 4.5 sits strictly between.
 - Exit-code convention: `0` clean; non-zero classified as failure. Do not interpret stderr for pass/fail.
 
-For verify-side severity classification (when this gate's `status="fail"` reaches the verify entry gate), see `./threat-modeling.md` — STRIDE dispositions are the audit-side framework that informs whether a failed quality gate blocks ship.
+For verify-side severity classification (when this gate's `status="fail"` reaches the verify entry gate), see `./threat-modeling.md` - STRIDE dispositions are the audit-side framework that informs whether a failed quality gate blocks ship.

package/skills/quality-gate/threat-modeling.md

@@ -25,35 +25,35 @@ Apply STRIDE during:
 - **Audit pillar runs** when a heuristic flags potential security surface.
 - **Plan-phase risk-register population** when the plan touches user input, network
   endpoints, file IO from user paths, or persisted state.
-- **Threat register on plans that ship new endpoints** — assign one of {mitigate, accept,
+- **Threat register on plans that ship new endpoints** - assign one of {mitigate, accept,
   transfer} to every identified threat before the plan ships.
 
 ## STRIDE categories
 
 | Letter | Threat                  | Audit lens                                         |
 |--------|-------------------------|----------------------------------------------------|
-| S      | Spoofing                | Auth surfaces — login, session, token issuance     |
-| T      | Tampering               | Data integrity — write paths, persisted state      |
-| R      | Repudiation             | Audit trails / logging — proof of action           |
-| I      | Information Disclosure  | PII / secret leakage — logs, errors, side channels |
-| D      | Denial of Service       | Resource exhaustion — unbounded loops, large reads |
-| E      | Elevation of Privilege  | AuthZ bypass — role checks, capability tokens      |
+| S      | Spoofing                | Auth surfaces - login, session, token issuance     |
+| T      | Tampering               | Data integrity - write paths, persisted state      |
+| R      | Repudiation             | Audit trails / logging - proof of action           |
+| I      | Information Disclosure  | PII / secret leakage - logs, errors, side channels |
+| D      | Denial of Service       | Resource exhaustion - unbounded loops, large reads |
+| E      | Elevation of Privilege  | AuthZ bypass - role checks, capability tokens      |
 
 ## Trust boundaries
 
 A trust boundary is a point where untrusted input crosses into trusted code. Identify
-trust boundaries before applying STRIDE — each boundary is one analysis sweep.
+trust boundaries before applying STRIDE - each boundary is one analysis sweep.
 
 Identification heuristics:
 
-- **Network ingress** — HTTP, gRPC, WebSocket, MCP transport, any TCP/UDP listen socket.
-- **File reads from user-writable paths** — uploads, `$HOME` configs, user-supplied paths
+- **Network ingress** - HTTP, gRPC, WebSocket, MCP transport, any TCP/UDP listen socket.
+- **File reads from user-writable paths** - uploads, `$HOME` configs, user-supplied paths
   from CLI args, drag-drop.
-- **Subprocess spawns with user-supplied args** — `exec`/`spawn` where any argv element
+- **Subprocess spawns with user-supplied args** - `exec`/`spawn` where any argv element
   is reachable from user input (URL params, env vars, config keys).
-- **Deserialization of persisted format** — JSON, YAML, MsgPack, Protobuf, custom
+- **Deserialization of persisted format** - JSON, YAML, MsgPack, Protobuf, custom
   formats. The deserializer is the boundary, regardless of where the bytes came from.
-- **Third-party SDK callouts** — when gdd hands data to a peer-CLI, the data leaves the
+- **Third-party SDK callouts** - when gdd hands data to a peer-CLI, the data leaves the
   trust boundary; treat the return path as untrusted on re-entry.
 
 ## Disposition framework
@@ -69,7 +69,7 @@ Every identified threat MUST carry a disposition before the plan ships. Three va
 | Transfer    | Third-party owns the control surface (e.g., the OS, the runtime, a peer's   |
 |             | sandbox). Document the boundary; do not re-implement the control.            |
 
-Mitigations on Plan tasks are correctness requirements — the executor applies Rule 2
+Mitigations on Plan tasks are correctness requirements - the executor applies Rule 2
 (missing critical functionality) if a mitigation disposition is present but the
 implementation lacks the control.
 
@@ -87,15 +87,15 @@ When a plan carries a `<threat_model>` block in its frontmatter, each entry foll
   rationale: "<why accept/transfer>"                    # required when accept/transfer
 ```
 
-Multiple threats per plan are normal. The disposition column is the load-bearing field —
+Multiple threats per plan are normal. The disposition column is the essential field -
 the executor scans it; the verifier scans it.
 
 ## Cross-references
 
-- `./audit-scoring.md` — design-side audit-scoring rubric; STRIDE is one of its lenses.
-- `./anti-patterns.md` — concrete anti-patterns mapped to STRIDE categories where
+- `./audit-scoring.md` - design-side audit-scoring rubric; STRIDE is one of its lenses.
+- `./anti-patterns.md` - concrete anti-patterns mapped to STRIDE categories where
   applicable (e.g., `eval`-on-user-input → Tampering + EoP).
-- `./accessibility.md` — accessibility is the orthogonal lens; threat-modeling does not
+- `./accessibility.md` - accessibility is the orthogonal lens; threat-modeling does not
   cover it.
-- ASVS (OWASP Application Security Verification Standard) — external authority for the
+- ASVS (OWASP Application Security Verification Standard) - external authority for the
   control catalog. Cited in plan threat-registers as `ASVS V<chapter>.<section>`.

package/skills/quick/SKILL.md

@@ -12,9 +12,9 @@ Fast pipeline run. Skips optional-quality agents for speed while keeping the cor
 
 ## Default skipped agents
 
-- `design-phase-researcher` — no external research step
-- `design-assumptions-analyzer` — no assumption surfacing
-- `design-integration-checker` — skipped (verifier still runs)
+- `design-phase-researcher` - no external research step
+- `design-assumptions-analyzer` - no assumption surfacing
+- `design-integration-checker` - skipped (verifier still runs)
 
 ## Default kept agents
 
@@ -38,7 +38,7 @@ Fast pipeline run. Skips optional-quality agents for speed while keeping the cor
 
 ## Do Not Use When
 
-- First pipeline run in a new project — use the full pipeline.
-- Large or cross-cutting changes — skip risks are higher.
+- First pipeline run in a new project - use the full pipeline.
+- Large or cross-cutting changes - skip risks are higher.
 
 ## QUICK COMPLETE

package/skills/reapply-patches/SKILL.md

@@ -12,16 +12,16 @@ Re-applies user customizations to `reference/*.md` files after `/gdd:update` res
 
 ## Steps
 
-1. **Detect baseline** — check for `.gdd-baseline/reference/`. If it does not exist, warn the user and offer to initialize it from the current `reference/` state (no diffs will be generated on this run; the next `/gdd:update` will use the new baseline).
-2. **Generate diffs** — for each `reference/*.md`, run `git diff --no-index .gdd-baseline/reference/<file> reference/<file>` (or plain `diff -u`). Non-empty diffs are candidate user patches.
-3. **Review patches** — for each non-empty diff, print the hunk and ask: `Apply this customization to the updated reference/<file>? (yes/no/view)`. `view` shows the full diff; `yes` queues the patch; `no` skips it.
-4. **Apply** — for each confirmed patch, apply it with `patch -p0` or with a targeted `Edit` call. Log every applied patch.
-5. **Refresh baseline** — after applying, copy the current `reference/*.md` tree into `.gdd-baseline/reference/` to record the new pristine-for-this-user state.
-6. **Dry-run** — if `--dry-run` is passed, perform steps 1–3 but do not apply patches or refresh the baseline.
+1. **Detect baseline** - check for `.gdd-baseline/reference/`. If it does not exist, warn the user and offer to initialize it from the current `reference/` state (no diffs will be generated on this run; the next `/gdd:update` will use the new baseline).
+2. **Generate diffs** - for each `reference/*.md`, run `git diff --no-index .gdd-baseline/reference/<file> reference/<file>` (or plain `diff -u`). Non-empty diffs are candidate user patches.
+3. **Review patches** - for each non-empty diff, print the hunk and ask: `Apply this customization to the updated reference/<file>? (yes/no/view)`. `view` shows the full diff; `yes` queues the patch; `no` skips it.
+4. **Apply** - for each confirmed patch, apply it with `patch -p0` or with a targeted `Edit` call. Log every applied patch.
+5. **Refresh baseline** - after applying, copy the current `reference/*.md` tree into `.gdd-baseline/reference/` to record the new pristine-for-this-user state.
+6. **Dry-run** - if `--dry-run` is passed, perform steps 1–3 but do not apply patches or refresh the baseline.
 
 ## Design note
 
-`.gdd-baseline/` is per-install and user-local. It must be gitignored by the host project — it is never committed to the plugin repo.
+`.gdd-baseline/` is per-install and user-local. It must be gitignored by the host project - it is never committed to the plugin repo.
 
 ## Output
 

package/skills/reflect/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: gdd-reflect
-description: "Run design-reflector on demand — produces .design/reflections/<cycle-slug>.md with improvement proposals. Review proposals with /gdd:apply-reflections."
+description: "Run design-reflector on demand - produces .design/reflections/<cycle-slug>.md with improvement proposals. Review proposals with /gdd:apply-reflections."
 argument-hint: "[--dry-run] [--cycle <slug>]"
 tools: Read, Write, Task
 ---
@@ -9,7 +9,7 @@ tools: Read, Write, Task
 
 # /gdd:reflect
 
-Run `design-reflector` on demand against the current (or specified) cycle. Produces `.design/reflections/<cycle-slug>.md` with numbered improvement proposals. Every proposal requires explicit user review — nothing is auto-applied.
+Run `design-reflector` on demand against the current (or specified) cycle. Produces `.design/reflections/<cycle-slug>.md` with numbered improvement proposals. Every proposal requires explicit user review - nothing is auto-applied.
 
 ## Steps
 
@@ -81,5 +81,5 @@ Run `design-reflector` on demand against the current (or specified) cycle. Produ
 
 - Do not auto-apply any proposal.
 - Do not modify agent files, reference files, or budget.json.
-- Do not run the full audit pipeline — this is a standalone reflection run.
+- Do not run the full audit pipeline - this is a standalone reflection run.
 - Do not bypass the threshold knob. The default `reflector.capability_gap_threshold: 3` exists to suppress noise; do NOT lower it below 1.

package/skills/reflect/procedures/capability-gap-scan.md

@@ -6,9 +6,9 @@ Run during the reflection pass to detect recurring patterns lacking a dedicated
 
 Three signal sources are scanned. All paths are repo-relative; the scan tolerates absent sources (returns `[]` from that source, no error):
 
-- `.design/intel/*.md` — Phase 11 intel-store slice files. The scan extracts `Touches:` lines, tokenizes the comma-separated value, and clusters slices that share the same canonical `(sortedTouches, agent_type)` signal.
-- `.design/telemetry/posterior.json` — Phase 23.5 bandit posterior file written by `scripts/lib/bandit-router.cjs`. The scan reads `arms[]` and flags arms whose `count >= threshold` AND whose `agent` is in `GENERIC_AGENT_FALLBACKS` (`general-purpose`, `default-executor`, `fallback`, `generic`) OR is not in the project's specialized-agent set.
-- `.design/gep/events.jsonl` — Phase 22 typed-causal event chain via `scripts/lib/event-chain.cjs`. The scan filters rows to the last `windowDays` (default 30), groups by `(sortedDecisionRefs, agent)`, and flags sequences that recur ≥ threshold times.
+- `.design/intel/*.md` - Phase 11 intel-store slice files. The scan extracts `Touches:` lines, tokenizes the comma-separated value, and clusters slices that share the same canonical `(sortedTouches, agent_type)` signal.
+- `.design/telemetry/posterior.json` - Phase 23.5 bandit posterior file written by `scripts/lib/bandit-router.cjs`. The scan reads `arms[]` and flags arms whose `count >= threshold` AND whose `agent` is in `GENERIC_AGENT_FALLBACKS` (`general-purpose`, `default-executor`, `fallback`, `generic`) OR is not in the project's specialized-agent set.
+- `.design/gep/events.jsonl` - Phase 22 typed-causal event chain via `scripts/lib/event-chain.cjs`. The scan filters rows to the last `windowDays` (default 30), groups by `(sortedDecisionRefs, agent)`, and flags sequences that recur ≥ threshold times.
 
 ## Outputs
 
@@ -29,11 +29,11 @@ The orchestrator `runCapabilityGapScan(opts)` in `scripts/lib/reflector/capabili
 
 Key entry points (all exported from the module):
 
-- `computeContextHash({touches, agent_type})` — pure deterministic hash (sha256 of normalized JSON; touches are sorted ASCII-asc).
+- `computeContextHash({touches, agent_type})` - pure deterministic hash (sha256 of normalized JSON; touches are sorted ASCII-asc).
 - `scanIntelTouchesClusters({intelDir, existingAgents, threshold, baseDir})`
 - `scanPosteriorArms({posteriorPath, specializedAgents, threshold, baseDir})`
 - `scanTrajectorySlices({chainPath, windowDays, threshold, specializedAgents, baseDir})`
-- `runCapabilityGapScan(opts)` — orchestrator.
+- `runCapabilityGapScan(opts)` - orchestrator.
 
 The `context_hash` is the join key for Plan 29-03's aggregation: the same signal across runs produces the same hash regardless of touches-list ordering.
 
@@ -82,7 +82,7 @@ interface TrajectoryRef {
 }
 ```
 
-Internally, the scan carries a line-based `{path, lineStart, lineEnd, sha256}` shape on each `Finding.evidence_refs`. The `lineRefToTrajectoryRef` translator converts to the schema shape at emit time. The sha256 algorithm: read lines `[lineStart..lineEnd]` (1-based inclusive), join with `'\n'` (no trailing newline — stable across OSes), sha256 the UTF-8 bytes.
+Internally, the scan carries a line-based `{path, lineStart, lineEnd, sha256}` shape on each `Finding.evidence_refs`. The `lineRefToTrajectoryRef` translator converts to the schema shape at emit time. The sha256 algorithm: read lines `[lineStart..lineEnd]` (1-based inclusive), join with `'\n'` (no trailing newline - stable across OSes), sha256 the UTF-8 bytes.
 
 Consumers (Plan 29-03 + audit tooling) re-read the slice and recompute the hash; mismatch = chain mutation; abort + warn.
 
@@ -98,7 +98,7 @@ Without `--dry-run`, the CLI writes events to the chain file and prints a one-li
 
 ## Testing
 
-Tests live at `tests/reflector-capability-gap.test.cjs` and run on synthetic in-tmpdir fixtures only (D-11 — no live writes to real `.design/`). Each test passes an injected `emit` spy so no real `appendChainEvent` calls occur. The hash-pin mutation-detection regression is enforced by a dedicated test: re-read the pointed-to slice, recompute the sha256, and assert mismatch after the source file is mutated.
+Tests live at `tests/reflector-capability-gap.test.cjs` and run on synthetic in-tmpdir fixtures only (D-11 - no live writes to real `.design/`). Each test passes an injected `emit` spy so no real `appendChainEvent` calls occur. The hash-pin mutation-detection regression is enforced by a dedicated test: re-read the pointed-to slice, recompute the sha256, and assert mismatch after the source file is mutated.
 
 Run the tests directly:
 
@@ -114,7 +114,7 @@ npm test
 
 ## See also
 
-- `reference/schemas/events.schema.json` — the `capability_gap` event class shipped by Plan 29-01 (the 7-field shape + `TrajectoryRef` definition).
-- `scripts/lib/event-chain.cjs` — `appendChainEvent` (the real emitter API; 29-01 did NOT ship a separate helper file).
-- `scripts/lib/bandit-router.cjs` — Phase 23.5 posterior file producer (the source for scan #2).
-- `.planning/phases/29-capability-gap-self-authoring/CONTEXT.md` — phase decisions (D-02 7-field shape, D-07 hash-pinning, D-08 MCP carve-out, D-11 tmpdir tests).
+- `reference/schemas/events.schema.json` - the `capability_gap` event class shipped by Plan 29-01 (the 7-field shape + `TrajectoryRef` definition).
+- `scripts/lib/event-chain.cjs` - `appendChainEvent` (the real emitter API; 29-01 did NOT ship a separate helper file).
+- `scripts/lib/bandit-router.cjs` - Phase 23.5 posterior file producer (the source for scan #2).
+- `.planning/phases/29-capability-gap-self-authoring/CONTEXT.md` - phase decisions (D-02 7-field shape, D-07 hash-pinning, D-08 MCP carve-out, D-11 tmpdir tests).

package/skills/report-issue/SKILL.md

@@ -11,14 +11,14 @@ Local-first, consent-gated GitHub issue reporter. No auto-mode. Destination repo
 
 ## Pre-flight
 
-If invoked without an error context (the user just typed `/gdd:report-issue` cold), ask in one round-trip: *"What command failed and what was the error?"* Capture command name + a paraphrase of the error. Don't interrogate — one short answer is enough to feed the triage matcher.
+If invoked without an error context (the user just typed `/gdd:report-issue` cold), ask in one round-trip: *"What command failed and what was the error?"* Capture command name + a paraphrase of the error. Don't interrogate - one short answer is enough to feed the triage matcher.
 
 ## Steps
 
 0. **Kill-switch** (D-08). Call `isDisabled()` from `scripts/lib/issue-reporter/kill-switch.cjs` FIRST. Either env (`GDD_DISABLE_ISSUE_REPORTER=1`) OR config (`.design/config.json` with `{ "issue_reporter": false }`) makes the command unavailable. Surface the disable line via `getDisableReason()` ('env' wins when both set) and stop. No draft, no triage, no payload. `gsd-health` mirrors the same line for at-a-glance verification.
 1. **Triage**. Call `matchKnownFailure(errorContext)` from `scripts/lib/issue-reporter/triage-matcher.cjs` (Plan 30-03). If `matched === true` and `--force-report` is NOT set: print the suggested diagnosis + remedy, stop. Do NOT write a draft. (D-07)
 2. **Assemble**. Call `assemble(commandName, errorContext, trajectoryRef?, capabilityGapEvent?)` from `scripts/lib/issue-reporter/payload-assembly.cjs` (Plan 30-02). This layers Phase 22 redact → Phase 30 pseudonymize, computes the fingerprint, and renders bilingual disclaimer + sections.
-3. **Draft**. Call `writeDraft({title, body, fingerprint})` from `scripts/lib/issue-reporter/draft-writer.cjs`. Print the absolute path: `Draft written to .design/issue-drafts/<timestamp>-<fp8>.md`. The file persists across decline — the user keeps their work either way. (D-04)
+3. **Draft**. Call `writeDraft({title, body, fingerprint})` from `scripts/lib/issue-reporter/draft-writer.cjs`. Print the absolute path: `Draft written to .design/issue-drafts/<timestamp>-<fp8>.md`. The file persists across decline - the user keeps their work either way. (D-04)
 4. **Dedup** (D-06). Call `searchByFingerprint(fingerprint, {destination})` from `scripts/lib/issue-reporter/dedup.cjs`. If `degraded === true` show a one-line warning and fall through to Step 5. If `matches.length === 0` fall through to Step 5 unchanged. If `matches.length >= 1` render the dedup UI listing each `{number,title,url}` with three actions per match: **`+1`** → `react(n, {destination})` (no new issue, D-06), **`me-too`** → `commentMeToo(n, {destination, errorContext, runtime, pluginVersion})` body is EXACTLY 3 fields (last error line, runtime, plugin version) from the ALREADY-pseudonymized 30-02 pipeline (D-01), **`new`** → fall through to Step 5 with the prepared draft despite the match.
 5. **Edit**. If `$EDITOR` is set, the consent prompt opens it on the draft and blocks until exit. Otherwise the path is printed and the user can open it manually. `EDITOR` is the only env var the report flow reads.
 6. **Consent**. Call `promptConsent({draftPath})` from `scripts/lib/issue-reporter/consent-prompt.cjs`. Re-reads the (possibly edited) draft from disk, prints a summary, asks `Submit this issue to hegemonart/get-design-done? [y/N]`. Anything other than `y`/`yes` declines. (D-03)
@@ -29,7 +29,7 @@ If invoked without an error context (the user just typed `/gdd:report-issue` col
 
 ## Use through `--report`
 
-Whitelisted to specific commands via `scripts/lib/issue-reporter/cli-flag-report.cjs`. Only `gdd:plan-phase`, `gdd:execute-phase`, and `gdd:report-issue` itself install the flag — and only when `reference/known-failure-modes.md` has at least one `propose_report: true` entry. Non-whitelisted commands silently do not see the flag. (D-11)
+Whitelisted to specific commands via `scripts/lib/issue-reporter/cli-flag-report.cjs`. Only `gdd:plan-phase`, `gdd:execute-phase`, and `gdd:report-issue` itself install the flag - and only when `reference/known-failure-modes.md` has at least one `propose_report: true` entry. Non-whitelisted commands silently do not see the flag. (D-11)
 
 ## Use through `--force-report`
 
@@ -44,10 +44,10 @@ Available on `/gdd:report-issue`. Overrides the triage gate (Step 1) but does NO
 
 ## Troubleshooting
 
-- **`gh` not authenticated.** The submission step fails with a clear error pointing to the draft path. Run `gh auth login` and re-invoke `/gdd:report-issue` — the draft survives.
+- **`gh` not authenticated.** The submission step fails with a clear error pointing to the draft path. Run `gh auth login` and re-invoke `/gdd:report-issue` - the draft survives.
 - **No `EDITOR` set.** The path is printed; open it in whatever editor you prefer, save, then return to the consent prompt.
 - **Triage matched something irrelevant.** Pass `--force-report` to bypass the gate. Consent is still required.
 - **`gh` not installed (D-10).** After consent, the payload is copied to your clipboard and an issue-template URL is printed. Paste into the URL to file manually. Install `gh` later and the existing draft can be re-submitted.
-- **Command unavailable / disabled (D-08).** Run `gsd-health`. The `issue reporter:` line shows the active disable surface — either `disabled by env (GDD_DISABLE_ISSUE_REPORTER=1)` or `disabled by config (.design/config.json: issue_reporter=false)`. Unset the env var or flip the config key to re-enable.
+- **Command unavailable / disabled (D-08).** Run `gsd-health`. The `issue reporter:` line shows the active disable surface - either `disabled by env (GDD_DISABLE_ISSUE_REPORTER=1)` or `disabled by config (.design/config.json: issue_reporter=false)`. Unset the env var or flip the config key to re-enable.
 
 See [report-issue-procedure.md](./report-issue-procedure.md) for the full procedure with rationale per decision (D-02, D-03, D-04, D-05, D-07, D-11).

package/skills/report-issue/report-issue-procedure.md

@@ -1,4 +1,4 @@
-# `/gdd:report-issue` — Full Procedure
+# `/gdd:report-issue` - Full Procedure
 
 Long-form companion to [SKILL.md](./SKILL.md). Phase 28.5 keeps SKILL.md ≤100 lines; step-by-step + rationale live here.
 
@@ -9,32 +9,32 @@ The report flow is the only outbound path the plugin offers. Every byte leaving
 1. has been redacted for secrets (Phase 22 `redact.cjs`),
 2. has been pseudonymized for identity (Plan 30-01 `pseudonymize.cjs`),
 3. has been written to disk where the user can read it,
-4. has cleared a pre-submit dedup check against the destination repo (Plan 30-05) — `+1` and `me-too` on a matching existing issue NEVER spawn a duplicate (D-06),
+4. has cleared a pre-submit dedup check against the destination repo (Plan 30-05) - `+1` and `me-too` on a matching existing issue NEVER spawn a duplicate (D-06),
 5. has been read back from disk after the user closed the editor, and
 6. has cleared an explicit per-issue `y/N` prompt.
 
 No environment variable, command-line flag, or build configuration bypasses any of these steps. Two test layers enforce this:
 
-- **Static** (`tests/report-issue-destination-static.test.cjs`, `tests/report-issue-no-auto-submit-static.test.cjs`) — fail the build on any forbidden code pattern in `skills/report-issue/` or `scripts/lib/issue-reporter/`.
-- **Runtime** (`tests/report-issue.test.cjs`) — 26 cases proving the orchestrator threads consent + persistence + edit-before-submit + triage + hardcoded destination + flag whitelist.
+- **Static** (`tests/report-issue-destination-static.test.cjs`, `tests/report-issue-no-auto-submit-static.test.cjs`) - fail the build on any forbidden code pattern in `skills/report-issue/` or `scripts/lib/issue-reporter/`.
+- **Runtime** (`tests/report-issue.test.cjs`) - 26 cases proving the orchestrator threads consent + persistence + edit-before-submit + triage + hardcoded destination + flag whitelist.
 
 ## Steps
 
-### Step 1 — Triage gate (D-07)
+### Step 1 - Triage gate (D-07)
 
 `matchKnownFailure(errorContext)` regex-matches `error.message + error.stack` against `reference/known-failure-modes.md`. If matched, prints diagnosis + remedy and exits without writing a draft. `--force-report` overrides the gate but does NOT override consent.
 
-### Step 2 — Assemble payload (D-01)
+### Step 2 - Assemble payload (D-01)
 
 `assemble(commandName, errorContext, trajectoryRef?, capabilityGapEvent?)` returns markdown. Order is locked: redact → pseudonymize (Case 9 of 30-02 enforces). Bilingual disclaimer at top: "Это псевдонимизация, не анонимизация" / "This is pseudonymization, not anonymization." Fingerprint computed on the scrubbed stack so the same bug from different cwd's hashes the same.
 
-### Step 3 — Write draft (D-04)
+### Step 3 - Write draft (D-04)
 
-`writeDraft({title, body, fingerprint})` writes to `.design/issue-drafts/<YYYYMMDDTHHMMSSZ>-<fp8>.md`. The file has a small HTML-comment header (timestamp, destination, full fingerprint) so a future maintainer with a corrupted-looking draft can reconstruct provenance. The file is NOT deleted on decline — the user keeps their work.
+`writeDraft({title, body, fingerprint})` writes to `.design/issue-drafts/<YYYYMMDDTHHMMSSZ>-<fp8>.md`. The file has a small HTML-comment header (timestamp, destination, full fingerprint) so a future maintainer with a corrupted-looking draft can reconstruct provenance. The file is NOT deleted on decline - the user keeps their work.
 
-### Step 4 — Pre-submit dedup (D-06)
+### Step 4 - Pre-submit dedup (D-06)
 
-`scripts/lib/issue-reporter/dedup.cjs` exports `searchByFingerprint(fingerprint, {destination})` which spawns `gh issue list --search "fingerprint:<hash>" --json number,title,url --repo <destination>` (read-only). Resolves `{matches: [...], degraded?, reason?}` — NEVER throws on gh failure.
+`scripts/lib/issue-reporter/dedup.cjs` exports `searchByFingerprint(fingerprint, {destination})` which spawns `gh issue list --search "fingerprint:<hash>" --json number,title,url --repo <destination>` (read-only). Resolves `{matches: [...], degraded?, reason?}` - NEVER throws on gh failure.
 
 Routing:
 
@@ -42,18 +42,18 @@ Routing:
 - `matches.length >= 1` → render the dedup UI listing each `{number, title, url}` with three actions per match:
 
   - **`+1`** → `react(n, {destination})` spawns `gh api -X POST /repos/<destination>/issues/<n>/reactions -f content=+1`. Resolves `{ok:true}`; exits the report flow on success ("reaction recorded on #<n>"). **NO new issue is created** (D-06).
-  - **`me-too`** → `commentMeToo(n, {destination, errorContext, runtime, pluginVersion})` spawns `gh issue comment <n> --repo <destination> --body <body>`. The body contains EXACTLY three fields (`Last error:`, `Runtime:`, `Plugin version:`) — nothing more (negative-presence tested). `errorContext.lastErrorLine` is the ALREADY-pseudonymized last error line from 30-02's payload pipeline (D-01); dedup.cjs does NOT re-derive raw stderr. Exits the report flow on success ("comment added to #<n>"). **NO new issue is created** (D-06).
+  - **`me-too`** → `commentMeToo(n, {destination, errorContext, runtime, pluginVersion})` spawns `gh issue comment <n> --repo <destination> --body <body>`. The body contains EXACTLY three fields (`Last error:`, `Runtime:`, `Plugin version:`) - nothing more (negative-presence tested). `errorContext.lastErrorLine` is the ALREADY-pseudonymized last error line from 30-02's payload pipeline (D-01); dedup.cjs does NOT re-derive raw stderr. Exits the report flow on success ("comment added to #<n>"). **NO new issue is created** (D-06).
   - **`new`** → fall through to Step 5 with the prepared draft despite the match (user explicitly opted to force a new issue).
 
-`+1` and `me-too` failures (auth/network/rate) propagate as rejected promises with annotated `.reason` so the caller can offer retry/cancel — they do NOT silently fall back to creating a new issue (that would defeat dedup intent).
+`+1` and `me-too` failures (auth/network/rate) propagate as rejected promises with annotated `.reason` so the caller can offer retry/cancel - they do NOT silently fall back to creating a new issue (that would defeat dedup intent).
 
 Wiring: `runReportFlow` calls `options.dedupCheck({fingerprint, title})` BEFORE the consent prompt (`report-flow.cjs` STEP 4). The skill drives the `+1`/`me-too`/`new` UI by passing a `dedupCheck` callback that wraps `searchByFingerprint` + the `react`/`commentMeToo` calls. Returning truthy `existing` from the callback short-circuits `runReportFlow` to `{submitted:false, reason:'duplicate'}`. Returning falsy continues to Step 5.
 
-### Step 5 — Edit (optional)
+### Step 5 - Edit (optional)
 
 If `$EDITOR` is set, `promptConsent` spawns it on the draft path and blocks until exit. Otherwise the user opens it manually. `EDITOR` is a POSIX convention (git, crontab, gh all use it); the static-grep test only forbids env-var reads matching `/REPORT|ISSUE|AUTO_REPORT/i`.
 
-### Step 6 — Consent prompt (D-03)
+### Step 6 - Consent prompt (D-03)
 
 The single submission gate for the new-issue path. Three preconditions must hold:
 
@@ -63,9 +63,9 @@ The single submission gate for the new-issue path. Three preconditions must hold
 
 The function prints a summary (destination, draft path, title, first 10 body lines), asks `Submit this issue to hegemonart/get-design-done? [y/N]` via `readline`, treats anything other than `y`/`yes` (case-insensitive, trimmed) as decline, and **re-reads the draft from disk** so user edits in Step 5 are picked up.
 
-(The `+1` / `me-too` paths from Step 4 do NOT pass through this consent prompt — selecting either action in the dedup UI IS the explicit consent for that minimal interaction. The new-issue path always passes through this prompt.)
+(The `+1` / `me-too` paths from Step 4 do NOT pass through this consent prompt - selecting either action in the dedup UI IS the explicit consent for that minimal interaction. The new-issue path always passes through this prompt.)
 
-### Step 7 — Submit via `gh` (D-05 + D-02)
+### Step 7 - Submit via `gh` (D-05 + D-02)
 
 `submitViaGh({title, body})` spawns:
 
@@ -73,7 +73,7 @@ The function prints a summary (destination, draft path, title, first 10 body lin
 gh issue create --repo hegemonart/get-design-done --title <title> --body-file <tmp/body.md>
 ```
 
-Body is written to a tmp file to avoid arg-length and shell-escaping. URL parsed from stdout. No HTTPS, no fetch, no third-party packages — only the user's `gh` CLI with their credentials.
+Body is written to a tmp file to avoid arg-length and shell-escaping. URL parsed from stdout. No HTTPS, no fetch, no third-party packages - only the user's `gh` CLI with their credentials.
 
 ## The `--report` flag (D-11)
 
@@ -92,8 +92,8 @@ Body is written to a tmp file to avoid arg-length and shell-escaping. URL parsed
 | Runtime consent | Only `y`/`yes` accepted | C1..C3, U3 |
 | Runtime re-read | `promptConsent` re-reads draft before returning final body | E1, E2 |
 | Runtime destination | `submitViaGh` always passes `--repo hegemonart/get-design-done` | H1 |
-| Dedup destination | `dedup.cjs` accepts `destination` only as a parameter — no env/config lookup (D-02) | `issue-reporter-dedup.test.cjs` test 11 |
-| Dedup body shape | `buildMeTooBody` returns EXACTLY 3 lines (`Last error:` / `Runtime:` / `Plugin version:`) — no stack/path/env/cmd (D-06) | tests 5 + 6 (verbatim + negative-presence) |
+| Dedup destination | `dedup.cjs` accepts `destination` only as a parameter - no env/config lookup (D-02) | `issue-reporter-dedup.test.cjs` test 11 |
+| Dedup body shape | `buildMeTooBody` returns EXACTLY 3 lines (`Last error:` / `Runtime:` / `Plugin version:`) - no stack/path/env/cmd (D-06) | tests 5 + 6 (verbatim + negative-presence) |
 | Dedup network | `dedup.cjs` imports only `child_process`; no outbound URL literals, no global fetch primitive, no third-party HTTP client libraries (D-05) | 30-07 static-grep gate |
 | Dedup test hermeticity | No live `gh` calls in CI; injected spawn spy + traced `child_process.spawnSync` counter assert 0 real invocations (D-13) | `issue-reporter-dedup.test.cjs` test 10 |
 | Dedup pseudonymization | `me-too` body uses the ALREADY-pseudonymized `errorContext.lastErrorLine` from 30-02's pipeline; dedup.cjs does NOT re-derive raw stderr (D-01) | dedup test "commentMeToo passes pseudonymized lastErrorLine through to gh body" |
@@ -103,18 +103,18 @@ Body is written to a tmp file to avoid arg-length and shell-escaping. URL parsed
 - **`gh` not authenticated**: submission throws with status + stderr; draft path preserved. Run `gh auth login`, retry. (T-30-04-08 accepted)
 - **`EDITOR` spawns wrong tool**: set `EDITOR=<your-editor>` in shell rc.
 - **Triage matched something irrelevant**: pass `--force-report`. Consent still required.
-- **TTY refused (CI / non-interactive)**: by design — run locally. (T-30-04-05 mitigated)
+- **TTY refused (CI / non-interactive)**: by design - run locally. (T-30-04-05 mitigated)
 - **No `--report` flag on a command you expected**: not on the whitelist; file an issue via this flow describing the use case.
 
 ## Forward-looking hooks
 
-- **Plan 30-05** *(landed)* — `scripts/lib/issue-reporter/dedup.cjs` wires `options.dedupCheck` to `gh issue list --search "fingerprint:<hash>"`. The skill drives the `+1` / `me-too` / `new` UI by passing a `dedupCheck` callback that wraps `searchByFingerprint` + `react` + `commentMeToo`. The hook in `runReportFlow` now runs BEFORE consent (per D-06).
-- **Plan 30-06** *(landed)* — `scripts/lib/issue-reporter/kill-switch.cjs` adds the D-08 dual-surface disable check (`GDD_DISABLE_ISSUE_REPORTER=1` env OR `.design/config.json: { "issue_reporter": false }` config). `runReportFlow` calls `isDisabled()` as **Step 0**, BEFORE triage and any draft writing — when disabled, returns `{ submitted:false, reason:'disabled', surface:'env'|'config', message }`. `scripts/lib/issue-reporter/gh-absent-fallback.cjs` adds the D-10 clipboard+URL path (pbcopy / wl-copy → xclip / clip.exe) — invoked after consent when `detectGh()` returns false. `scripts/lib/gsd-health-mirror/index.cjs` mirrors the disable surface as a 5th health check; output is exactly one of: `issue reporter: enabled` / `issue reporter: disabled by env (GDD_DISABLE_ISSUE_REPORTER=1)` / `issue reporter: disabled by config (.design/config.json: issue_reporter=false)`. Tests at `tests/issue-reporter-fallback.test.cjs` (15 cases). Env wins precedence when both flags are set (matches gsd-health display).
-- **Plan 30-07** ships the network-isolation CI gate. Plans 30-04 and 30-05 already meet the invariant (no outbound URL literals, no global fetch primitive, no third-party HTTP client libraries — see `tests/issue-reporter-network-isolation.test.cjs` for the enforced forbidden-token list); the gate locks it in.
+- **Plan 30-05** *(landed)* - `scripts/lib/issue-reporter/dedup.cjs` wires `options.dedupCheck` to `gh issue list --search "fingerprint:<hash>"`. The skill drives the `+1` / `me-too` / `new` UI by passing a `dedupCheck` callback that wraps `searchByFingerprint` + `react` + `commentMeToo`. The hook in `runReportFlow` now runs BEFORE consent (per D-06).
+- **Plan 30-06** *(landed)* - `scripts/lib/issue-reporter/kill-switch.cjs` adds the D-08 dual-surface disable check (`GDD_DISABLE_ISSUE_REPORTER=1` env OR `.design/config.json: { "issue_reporter": false }` config). `runReportFlow` calls `isDisabled()` as **Step 0**, BEFORE triage and any draft writing - when disabled, returns `{ submitted:false, reason:'disabled', surface:'env'|'config', message }`. `scripts/lib/issue-reporter/gh-absent-fallback.cjs` adds the D-10 clipboard+URL path (pbcopy / wl-copy → xclip / clip.exe) - invoked after consent when `detectGh()` returns false. `scripts/lib/gsd-health-mirror/index.cjs` mirrors the disable surface as a 5th health check; output is exactly one of: `issue reporter: enabled` / `issue reporter: disabled by env (GDD_DISABLE_ISSUE_REPORTER=1)` / `issue reporter: disabled by config (.design/config.json: issue_reporter=false)`. Tests at `tests/issue-reporter-fallback.test.cjs` (15 cases). Env wins precedence when both flags are set (matches gsd-health display).
+- **Plan 30-07** ships the network-isolation CI gate. Plans 30-04 and 30-05 already meet the invariant (no outbound URL literals, no global fetch primitive, no third-party HTTP client libraries - see `tests/issue-reporter-network-isolation.test.cjs` for the enforced forbidden-token list); the gate locks it in.
 
 ## References
 
-- [SKILL.md](./SKILL.md) — entry contract.
-- `reference/pseudonymization-rules.md` — full R1..R8 rule catalog (Plan 30-01).
-- `reference/known-failure-modes.md` — triage catalogue (Plan 30-03).
-- `.planning/phases/30-issue-reporter/CONTEXT.md` — phase decisions D-01..D-15.
+- [SKILL.md](./SKILL.md) - entry contract.
+- `reference/pseudonymization-rules.md` - full R1..R8 rule catalog (Plan 30-01).
+- `reference/known-failure-modes.md` - triage catalogue (Plan 30-03).
+- `.planning/phases/30-issue-reporter/CONTEXT.md` - phase decisions D-01..D-15.

package/skills/resume/SKILL.md

@@ -13,9 +13,9 @@ disable-model-invocation: true
 
 Inverse of `/gdd:pause`. Reads a checkpoint file, prints a clear "you were here" summary, and routes to the next command.
 
-## Step 0 — Prime cycle context
+## Step 0 - Prime cycle context
 
-Two paths — MCP preferred when available, file-read fallback otherwise. This runs BEFORE checkpoint restoration so the "you were here" summary has full cycle context (phase, plans, decisions).
+Two paths - MCP preferred when available, file-read fallback otherwise. This runs BEFORE checkpoint restoration so the "you were here" summary has full cycle context (phase, plans, decisions).
 
 ### MCP path (preferred)
 
@@ -26,7 +26,7 @@ When `mcp__gdd_status` is exposed (Phase 27.7+, registered via `npx @hegemonart/
 3. Call `mcp__gdd_decisions_list` (no args) → full D-XX list with rationale. Use for the "decisions you made" line in the resume summary.
 4. (Optional) Call `mcp__gdd_plans_list` (no args) → current phase plans + status, to identify next incomplete plan.
 
-Three to four MCP calls = full resume priming (~5s, ~32k tokens — Storybloq benchmark). Proceed to Step 1.
+Three to four MCP calls = full resume priming (~5s, ~32k tokens - Storybloq benchmark). Proceed to Step 1.
 
 ### File-read path (fallback)
 
@@ -36,7 +36,7 @@ When MCP tools are not available:
 2. Also `Read .design/CYCLES.md` (if present) to see prior cycle state for the recap.
 3. Proceed to Step 1.
 
-This path loads the same context in 3–5 file reads (~60s, ~46.5k tokens — file-reading baseline).
+This path loads the same context in 3–5 file reads (~60s, ~46.5k tokens - file-reading baseline).
 
 ## Steps
 
@@ -55,7 +55,7 @@ This path loads the same context in 3–5 file reads (~60s, ~46.5k tokens — fi
 
 3. **Read checkpoint**: load `.design/checkpoints/NN-*.md`. If not found, try `.design/HANDOFF.md` as legacy fallback.
 
-4. **Check paused status via MCP**: call `mcp__gdd_state__get` and inspect `status`. If it does **not** start with `paused:`, print "No pause to resume" and exit — the prior session was not paused via `/gdd:pause`, so there is nothing to restore.
+4. **Check paused status via MCP**: call `mcp__gdd_state__get` and inspect `status`. If it does **not** start with `paused:`, print "No pause to resume" and exit - the prior session was not paused via `/gdd:pause`, so there is nothing to restore.
 
 5. **Restore prior status via MCP**: parse the prior status from the `paused:<prior>` prefix. Call `mcp__gdd_state__set_status` with `status: <prior>` to restore the pre-pause state.
 
@@ -74,7 +74,7 @@ This path loads the same context in 3–5 file reads (~60s, ~46.5k tokens — fi
    ```
 
 9. **Staleness check**: compare mtime of `.design/` artifacts vs `src/` via Bash `stat` when available. If `src/` has commits newer than the checkpoint timestamp, warn:
-   "Source has changed since checkpoint NN — consider re-running explore or verify."
+   "Source has changed since checkpoint NN - consider re-running explore or verify."
 
 10. **Route recommendation** based on checkpoint `Stage:` field:
     - `brief` → "Run `/gdd:brief`"
@@ -86,8 +86,8 @@ This path loads the same context in 3–5 file reads (~60s, ~46.5k tokens — fi
 ## Do Not
 
 - Do not delete checkpoint files.
-- Do not mutate STATE.md directly — all STATE.md writes go through the `gdd-state` MCP tools above.
-- Do not auto-execute the next command — just recommend.
-- Do not call `mcp__gdd_state__transition_stage` — resume restores prior status without moving the pipeline.
+- Do not mutate STATE.md directly - all STATE.md writes go through the `gdd-state` MCP tools above.
+- Do not auto-execute the next command - just recommend.
+- Do not call `mcp__gdd_state__transition_stage` - resume restores prior status without moving the pipeline.
 
 ## RESUME COMPLETE

package/skills/review-backlog/SKILL.md

@@ -9,11 +9,11 @@ disable-model-invocation: true
 
 **Role:** Walk through parked backlog items and for each ask: promote to this cycle, keep parked, or archive.
 
-## Step 1 — Read backlog
+## Step 1 - Read backlog
 
 Read `.design/backlog/BACKLOG.md`. Parse each `## <title>` block with its metadata and body. If no parked items, print "No parked backlog items." and stop.
 
-## Step 2 — Loop
+## Step 2 - Loop
 
 For each item with `**Status**: parked`:
 
@@ -26,7 +26,7 @@ Added: <date>
 Promote to this cycle | Keep parked | Archive
 ```
 
-## Step 3 — Apply decision
+## Step 3 - Apply decision
 
 - **Promote**: append `- [ ] [YYYY-MM-DD] <title>` under `## P1 — High` in `.design/TODO.md` (create file from the TODO.md skeleton if missing). Update backlog item status to `**Status**: promoted` + `**Promoted**: YYYY-MM-DD`.
 - **Keep parked**: leave unchanged.

package/skills/review-decisions/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: gdd-review-decisions
-description: "Surfaces the async decision-review queue for team mode. Reads .design/reviews/<decision-id>/ entries and reports each decision's state in the proposed → reviewing → approved → locked machine (via scripts/lib/collab/review-queue.cjs), so a team can see what's awaiting review, what's approved, and what's locked. --pending shows only decisions still needing action. Read-only — it reports the queue; it never advances a decision (that's a reviewer's explicit call). Use to run an async design-decision review without a meeting."
+description: "Surfaces the async decision-review queue for team mode. Reads .design/reviews/<decision-id>/ entries and reports each decision's state in the proposed → reviewing → approved → locked machine (via scripts/lib/collab/review-queue.cjs), so a team can see what's awaiting review, what's approved, and what's locked. --pending shows only decisions still needing action. Read-only - it reports the queue; it never advances a decision (that's a reviewer's explicit call). Use to run an async design-decision review without a meeting."
 argument-hint: "[<decision-id>] [--pending]"
 user-invocable: true
 tools: Read, Bash, Grep, Glob
@@ -10,7 +10,7 @@ tools: Read, Bash, Grep, Glob
 
 Closes the async-review gap for team mode: design decisions move through an explicit review queue
 instead of being decided in a single operator's head. This skill reports where each decision is.
-**Read-only** — it surfaces state; advancing a decision is a reviewer's explicit action. Contract:
+**Read-only** - it surfaces state; advancing a decision is a reviewer's explicit action. Contract:
 `../../reference/multi-author-model.md`.
 
 ## Invocation
@@ -30,7 +30,7 @@ instead of being decided in a single operator's head. This skill reports where e
 3. **Render** grouped by state: `proposed` / `reviewing` / `approved` / `locked`, each listing the
    decision id + a one-line summary. For `--pending`, use `review-queue.pending(entries)` to show only
    non-locked ones. For a single `<decision-id>`, also print its audit trail (transitions + approvers).
-4. **Do not advance.** Reporting only — moving a decision forward (or `/gdd:unlock-decision`) is the
+4. **Do not advance.** Reporting only - moving a decision forward (or `/gdd:unlock-decision`) is the
    reviewer's explicit call.
 
 ## Output