skillpacks 0.1.0

package/packs/docs-health/claude/reconcile-dev-docs/archive/v0.0/SKILL.md

@@ -0,0 +1,182 @@
+---
+name: reconcile-dev-docs
+description: Reconcile development docs by auditing roadmap, todo, history, phase archives, specs, git history, and code reality
+type: analysis
+version: v0.0
+argument-hint: "[audit|fix] [tasks|specs|all]"
+---
+
+# Reconcile Dev Docs
+
+Audit or repair development documentation so the roadmap, current work, history, specs, and evidence from git/code all tell the same story.
+
+## Process
+
+### 1. Determine Mode and Scope
+
+Parse `$ARGUMENTS`:
+
+- **Mode**: `audit` (default, read-only) or `fix` (update development docs only)
+- **Scope**: `all` (default), `tasks`, or `specs`
+
+Resolve the canonical development docs:
+
+- `tasks/roadmap.md`
+- `tasks/todo.md`
+- `tasks/manual-todo.md` (if it exists)
+- `tasks/record-todo.md` (if it exists)
+- `tasks/recurring-todo.md` (if it exists)
+- `tasks/history.md`
+- `tasks/phases/*.md`
+- `tasks/ideas.md` (if it exists)
+- `specs/*.md` and `specs/*/*.md`
+- `docs/specifications/*.md` (fallback specs location)
+
+If no task docs exist, report that there is nothing to reconcile and recommend `/roadmap` or `/plan-phase` based on available specs.
+
+### 2. Gather Evidence
+
+Read the development docs and collect:
+
+- Roadmap phases, milestone checkboxes, acceptance criteria, manual tasks, record tasks, recurring tasks, and deferred items
+- Current todo phase, checked/unchecked steps, blockers, and review/results sections
+- Manual task blockers and `_(blocks: ...)_` / `_(after: ...)_` annotations
+- Record task conditions, non-blocking reasons, evidence paths, and promotion rules
+- Recurring task cadence, last run, next due, evidence paths, and escalation conditions
+- History entries with dates, completed phases/steps, commit SHAs, and claimed outcomes
+- Phase archives and their completion summaries
+- Spec titles, statuses, acceptance criteria, and major implementation claims
+
+Read git/code evidence:
+
+- `git status --short`
+- `git log --oneline -50`
+- `git log --name-only --since=<latest relevant doc date>` when timestamps suggest docs may be stale
+- Targeted file existence and `rg` checks for claims marked complete
+
+### 3. Reconcile Rules
+
+Classify findings:
+
+| Severity | Examples |
+|----------|----------|
+| **Error** | Docs contradict each other: roadmap marks a phase done while todo has unchecked required steps; history claims completion but no matching roadmap/todo/archive/git evidence exists |
+| **Warning** | Likely drift: recent commits changed areas covered by specs but history/todo were not updated; completed phase has no archive; manual blockers remain for completed steps |
+| **Info** | Cleanup opportunities: stale handoff notes, missing links between history entries and commits, specs that should be checked by `/spec-drift` |
+
+Use these specific checks:
+
+- `tasks/todo.md` checked steps should be reflected in `tasks/history.md` after shipping.
+- Completed roadmap phases should have corresponding `tasks/phases/phase-N.md` archives.
+- The active `tasks/todo.md` phase should match the first incomplete roadmap phase.
+- `tasks/manual-todo.md` should not contain unchecked blockers for completed todo steps unless the user explicitly overrode them.
+- `tasks/todo.md` should not contain condition-gated baseline measurements or future records unless they are current execution work; move clear cases to `tasks/record-todo.md`.
+- `tasks/manual-todo.md` should not contain non-blocking records unless they require human-only external action tied to a `blocks` or `after` step; move clear cases to `tasks/record-todo.md`.
+- `tasks/manual-todo.md` should not contain agent-executable work such as repo edits, SDK wiring, generated assets, local commands, tests, audits, or authenticated CLI/API operations; move clear cases to `tasks/todo.md`.
+- `tasks/recurring-todo.md` due items should remain advisory unless explicitly promoted into `tasks/todo.md`.
+- Recent commits that complete user-facing work should have a matching history entry.
+- Specs whose described areas changed after their last modification date should be flagged for `/spec-drift`.
+- Claimed completed work should have either git evidence, file evidence, test evidence, or an explicit note explaining why evidence is unavailable.
+
+### 4. Fix Mode
+
+In `fix` mode, apply only unambiguous development-doc changes:
+
+- Append missing factual entries to `tasks/history.md`.
+- Check roadmap milestones only when todo/archive/git evidence all support completion.
+- Create missing `tasks/phases/phase-N.md` archives from completed `tasks/todo.md` content.
+- Move the next roadmap phase into `tasks/todo.md` only when the current phase is clearly complete.
+- Move clearly misclassified non-blocking condition-gated records from `tasks/todo.md` or `tasks/manual-todo.md` into `tasks/record-todo.md`.
+- Move clearly misclassified cadence-based obligations from `tasks/todo.md` or `tasks/manual-todo.md` into `tasks/recurring-todo.md`.
+- Move clearly misclassified agent-executable work from `tasks/manual-todo.md` into `tasks/todo.md` with enough file/command detail for `/run`.
+- Add unresolved contradictions to `tasks/todo.md` under `## Development Docs Reconciliation`.
+- Archive existing `specs/` or `docs/specifications/` files before replacing or substantively rewriting them.
+- Write `tasks/reconciliation-report.md` with resolved, deferred, and remaining findings.
+
+Ask before applying any ambiguous change. If already in Plan mode and there are 2-3 concrete choices, prefer `request_user_input`; otherwise ask plainly in chat.
+
+## Output Format
+
+### Audit Mode
+
+```
+## Dev Docs Reconciliation
+
+### Errors (N)
+- **tasks/todo.md** - [problem]. Evidence: [roadmap/history/git reference].
+
+### Warnings (N)
+- **specs/foo.md** - [problem]. Recommended action: /spec-drift.
+
+### Info (N)
+- **tasks/history.md** - [cleanup opportunity].
+
+### Summary
+- Roadmap/todo alignment: [ok|issues]
+- History coverage: [ok|issues]
+- Phase archives: [ok|issues]
+- Spec freshness: [ok|issues]
+- Recommended next action: [specific skill or fix plan]
+```
+
+### Fix Mode
+
+Report the same sections plus:
+
+```
+### Fixed
+- [x] [file] - [change made]
+
+### Deferred
+- [ ] [finding requiring user judgment]
+```
+
+## Constraints
+
+- Read-only by default.
+- In `fix` mode, modify only development docs under `tasks/`, `specs/`, or `docs/specifications/`; archive existing spec documents before replacement per the Archive-First Replacement Policy.
+- Do not modify code, research docs, kanban cards, git history, or deployment state.
+- Do not rewrite or delete old `tasks/history.md` entries; append corrections or superseding notes.
+- Do not uncheck completed work automatically. Report unsupported completion claims unless the user approves the correction.
+- Treat `tasks/roadmap.md` as the strategic source of truth, `tasks/todo.md` as the active execution contract, `tasks/manual-todo.md` as human-only external step-linked work, `tasks/record-todo.md` and `tasks/recurring-todo.md` as advisory surfaces, and `tasks/history.md` as append-only evidence.
+
+## Alignment Page
+
+When this skill produces durable deliverables (research, specs, plans, reports, prototypes, or any document output), build a full-depth HTML alignment page at `alignment/reconcile-dev-docs-{topic}.html`. Use a normalized topic slug derived from the app, feature, research subject, report subject, or output filename.
+
+**Full content requirement.** The alignment page must contain the complete content of every proposed markdown deliverable -- every section, every finding, every detail, every list item. It is a thorough interactive review document, not a summary. Render the full deliverable content in clean, readable HTML with appropriate hierarchy, styling, and navigation. If the skill writes multiple scoped deliverables in one run, build one alignment page that contains all deliverables with anchor-linked navigation. Durable tracker artifacts, such as `research/assumption-tracker.md`, remain canonical markdown outputs but must also be fully rendered into the alignment page before approval.
+
+**Dark-mode styling.** Use a dark color scheme by default. Base CSS variables: `--bg: #0d1117; --surface: #161b22; --border: #30363d; --text: #c9d1d9; --text-muted: #8b949e; --accent: #58a6ff; --green: #3fb950; --red: #f85149; --orange: #d29922; --purple: #bc8cff;`. Apply `background: var(--bg); color: var(--text);` on body. Use `--surface` for cards, nav, and table headers. Use `--border` for all borders. Use `--purple` for question blocks and gate headings. Use `--accent` for links and section headings. Keep headings `color: #fff` or `var(--accent)` for hierarchy. Question block backgrounds should use `#1c2333`.
+
+**Alignment gates.** Treat gates as explicit review sections inside the HTML page. A gate blocks finalization until its required inline questions are answered and compiled into YAML. Include every gate that applies to the skill output, and include these gate types whenever relevant: evidence coverage, assumptions/confidence, scope/non-goals, candidate/verdict decisions, artifact destination, proposed file changes, coverage checkpoint, and post-approval route.
+
+**Report-only research gates.** For report-only or pre-approval research skills, the alignment page must explicitly contain evidence coverage, assumptions/confidence, recommended path, proposed file changes, and approval gates before any canonical research, spec, or task file is created or updated.
+
+
+**Required inline questions.** Each gate must contain at least one required inline question placed directly under the content it governs, inside a visually distinct question block. Each question must use radio-button inputs and include two standing options after the skill-generated choices: "Other / None of the above" backed by a multi-line text box for free-form input, and "Need clarification" backed by an optional notes box where the user can explain what is unclear. When any radio option other than "Other" or "Need clarification" is selected, show an optional "Additional notes" text box beneath it so the user can qualify their choice. Generate questions based on what genuinely needs user input -- do not add filler questions. Do not create a separate bottom "Decisions & Clarifications" section.
+
+**Gate YAML contract.** At the bottom of the page, include a "Compile Answers" button that aggregates answers from all inline gate questions throughout the page, including free-text notes. The button remains disabled until every required question has a selection, shows a count of remaining unanswered questions, and scrolls to the first unanswered question if clicked early. When every question is answered, generate a structured YAML block with one item per gate answer using this stable shape: `section`, `gate_type`, `status` (`answered`, `other`, or `needs-clarification`), `answer`, optional `notes`, and optional `target_artifact` or `target_path` when the gate controls file output. After successful compilation, automatically attempt to copy the YAML to the clipboard with the Clipboard API, display copy status, and display the YAML in a read-only textarea with an explicit "Copy YAML" button. The copy button must retry clipboard copy when supported and fall back to selecting the textarea contents when clipboard access is unavailable or blocked.
+
+**Pre-approval stop.** Before user approval, the next action is review of the HTML alignment page, not downstream routing. Ask the user to review the page and provide the compiled YAML answers. Do not include `Recommended next skill`, `Recommended next command`, or downstream routing language until after compiled YAML has been provided and the approved artifacts have been written or updated.
+
+**Diff highlighting on updates.** When the agent updates an existing alignment page after receiving compiled answers, highlight what changed since the previous version. The agent chooses inline annotation or side-by-side layout per situation.
+
+**Archiving.** Before replacing an existing alignment page, archive it to `docs/history/archive/YYYY-MM-DD/HHMMSS/alignment/reconcile-dev-docs-{topic}.html`.
+
+**Browser open.** Attempt to open the resulting HTML page in the browser and report whether the open succeeded or was blocked. A blocked browser-open attempt does not make the skill fail when the files were written correctly.
+
+## Archive-First Replacement Policy
+
+- Before replacing or substantively rewriting an existing canonical research/spec document (`research/**/*.md`, `specs/**/*.md`, or `docs/specifications/**/*.md`), copy the current file to `docs/history/archive/YYYY-MM-DD/HHMMSS/<original-relative-path>`.
+- Preserve the archived snapshot exactly as it existed before the change; do not edit the archived copy after creating it.
+- After the archive snapshot exists, write the updated document to the original canonical path.
+- Report both the archive path and the updated canonical path in the final output.
+- New files do not need archive snapshots. Append-only updates do not need archive snapshots unless an existing section is regenerated or rewritten.
+- Keep any existing user approval requirement before overwriting or replacing a document; archiving does not replace asking when the skill already requires approval.
+
+## Default Shipping Contract
+
+- **Default next-step routing:** when reporting completion, include either `Recommended next skill: <command>` or the two-line pair `**Next work:** <specific task or "none">` and `**Recommended next command:** <one command or route>` so the next operator has a concrete handoff.
+- If this skill creates or modifies tracked repository files, finish by committing and pushing all intended changes to the repository primary branch (`main` when present, otherwise `master`) before stopping, even if the user did not explicitly ask for commit/push.
+- Do not leave tracked changes or unpushed commits behind. If unrelated tracked work is already present, either include it in sensible commits too or stop and explain the blocker.
+- This contract does not override stricter safety rules about secrets, destructive history changes, release publication/tag confirmation, or production deploy confirmation.

package/packs/docs-health/codex/hygiene/SKILL.md

@@ -0,0 +1,225 @@
+---
+name: hygiene
+description: Audit project structure for convention violations, missing files, template drift, and cross-platform sync gaps; optionally auto-fix
+type: analysis
+version: v0.0
+argument-hint: "[audit|fix] [skills|tasks|docs|codex|all]"
+---
+
+# Hygiene - Project Structure Audit
+
+Checks that project files follow established conventions, flags gaps and drift, and optionally fixes mechanical issues. Think of it as a linter for project structure rather than code.
+
+## Process
+
+### 1. Determine Mode and Scope
+
+Parse `$ARGUMENTS`:
+
+- **Mode**: `audit` (default, read-only) or `fix` (apply mechanical fixes)
+- **Scope**: `skills`, `tasks`, `docs`, `codex`, or `all` (default)
+
+Read `CLAUDE.md` for project conventions. Read `docs/skills-reference.md` for the canonical skill list. For `docs` scope, also read `references/documentation-templates.md` from this skill directory.
+
+### 2. Audit Skills (`skills` scope)
+
+For each skill directory in `global/claude/*/SKILL.md`, `global/codex/*/SKILL.md`, `packs/*/claude/*/SKILL.md`, and `packs/*/codex/*/SKILL.md`:
+
+**Frontmatter check:**
+- Has `name` field — matches directory name
+- Has `description` field — non-empty, under 120 characters
+- Has `version` field — valid semver (X.Y.Z)
+- Has `argument-hint` field (can be empty)
+
+**Section check:**
+- Has a top-level `# Heading` that matches the skill's purpose
+- Has `## Process` section with numbered steps
+- Has `## Output Format` section (or explicit "no files written" note)
+- Has `## Constraints` section
+
+**Naming check:**
+- Directory name uses kebab-case
+- Directory name matches frontmatter `name`
+
+Flag each violation with the file path and what's wrong.
+
+### 3. Audit Tasks (`tasks` scope)
+
+Check expected project files exist based on project phase:
+
+**Always expected:**
+- `tasks/roadmap.md` — has content, not just a placeholder
+- `tasks/todo.md` — has content with checkable items (`- [ ]` or `- [x]`)
+- `tasks/manual-todo.md` — if it exists, has content with checkable human-only external items (`- [ ]` or `- [x]`) and `_(blocks: ...)_` or `_(after: ...)_` annotations
+- `tasks/record-todo.md` — if it exists, contains non-blocking condition-gated records with source, condition, non-blocking reason, evidence, and promotion rule fields
+- `tasks/recurring-todo.md` — if it exists, contains cadence-based tasks with cadence, owner/agent, last run, next due, evidence/output path, and escalation fields
+- `tasks/history.md` — exists
+
+**Expected if specs exist:**
+- `tasks/ideas.md` — exists if `specs/*.md` has 2+ files
+
+**Phase archives:**
+- If `tasks/roadmap.md` has completed phases (checked-off milestones), `tasks/phases/` should exist with corresponding `phase-N.md` files
+
+**Staleness (informational, not violations):**
+- `tasks/todo.md` has all items checked but phase isn't marked complete in `tasks/roadmap.md`
+- `tasks/manual-todo.md` has unchecked items that block completed steps in `tasks/todo.md`
+- `tasks/manual-todo.md` contains agent-executable work that belongs in `tasks/todo.md`
+- `tasks/record-todo.md` has eligible items that may need promotion into `tasks/todo.md`
+- `tasks/recurring-todo.md` has due items that may need promotion into `tasks/todo.md`
+- `tasks/roadmap.md` references phases that have no corresponding detail
+
+### 4. Audit Docs (`docs` scope)
+
+**Skills reference sync:**
+- Every global core directory has a corresponding entry in `docs/skills-reference.md`
+- Every pack directory is listed under its pack in `docs/skills-reference.md`
+- The global core table and pack skill lists are complete (no missing, no extras)
+
+**Required docs:**
+- `CLAUDE.md` exists at project root
+- If `AGENTS.md` exists, it should not contradict the workflow/pipeline conventions in `CLAUDE.md`
+
+**Documentation template audit:**
+- Classify generated Markdown files by path pattern and validate them against `references/documentation-templates.md`
+- Check canonical roots:
+  - `tasks/` for roadmap, todo, history, manual tasks, record tasks, recurring tasks, phase archives, handoff, deployment ledgers, and lessons
+  - `specs/` for implementation specifications and interview logs
+  - `research/` for research docs, interview logs, search logs, experiments, and reconciliation reports
+  - `docs/specifications/` only as a fallback spec location
+  - `alignment/*.html` for generated browser-review alignment artifacts from planning and prototype workflows
+- Check canonical root directories exist:
+  - `tasks/` — Warning if absent
+  - `research/` — Info if absent; Warning if absent AND research-pattern files (`icp-*.md`, `gtm-*.md`, `competitive-*.md`, `journey-*.md`, `metrics.md`, `monetization.md`, `customer-feedback*.md`) exist elsewhere (e.g., `docs/`)
+  - `specs/` — Info if absent; Warning if absent AND spec-pattern files exist elsewhere
+  - `alignment/` — Info if present; generated HTML review pages are allowed here and should not be treated as misplaced documentation
+- Flag legacy or drifted planning locations such as new `docs/plan.md` or new `docs/phases/` phase archives; current workflows write plans under `tasks/`
+- Exempt archived snapshots under `docs/history/archive/**` from template checks
+- Treat unknown hand-written docs as Info unless they are clearly generated workflow artifacts
+
+**Universal generated-doc checks:**
+- First non-empty Markdown heading is exactly one `#` H1
+- Required sections for the file family are present once and in the expected order
+- Non-ledger research/spec/task artifacts include a metadata quote block when the template requires it
+- Checkable workflow docs use `- [ ]` / `- [x]` items where later skills need task state
+- Main research/spec docs end with `## Next Steps`, or explicitly state why no next step exists
+
+**Family-specific checks:**
+- `tasks/roadmap.md` has a summary, phase overview, repeated `## Phase N:` sections, phase milestones or acceptance criteria, and cross-phase concerns when multi-phase
+- `tasks/todo.md` has checkable work, a priority task queue, or a priority documentation todo, and does not contain the full multi-phase roadmap except during legacy migration
+- `tasks/manual-todo.md` has checkable human-only external items, every unchecked item includes `_(blocks: Step N.X)_` or `_(after: Step N.X)_`, and no item is repo/code/config/test/audit/CLI/API work the agent can execute
+- `tasks/record-todo.md` has checkable non-blocking record items with source, condition, non-blocking reason, required data/access, measurement/query, target note, revisit cadence/date, completion evidence, and promotion rule
+- `tasks/recurring-todo.md` has cadence-based items with task, cadence, owner/agent, scope, trigger, last run, next due, command/skill, evidence/output path, and escalation conditions
+- `tasks/history.md` is append-only with dated entries
+- `tasks/phases/phase-N.md` has completed steps, milestone or acceptance criteria, and an `## On Completion` or equivalent completion summary
+- `specs/*.md`, `specs/{app}/*.md`, and fallback `docs/specifications/*.md` have spec sections for overview, goals, non-goals, detailed design, edge cases, test plan, acceptance criteria, and open questions or explicit `None`
+- `*-interview.md` logs include questions, options when presented, responses or decisions, and final deviations or coverage summary
+- `research/*.md` and `research/{app}/*.md` have metadata, summary, source/evidence orientation, assumptions or risks when relevant, and next steps
+- `research/*-search-log.md` files are clearly marked as supporting context and include queries, findings, and source attribution
+- `research/experiments/*.md` has hypothesis, method, success criteria, timeline, budget, decision rules, results, and next steps
+- `sync.md` uses `## Dependencies`, `## Conflict Resolution`, `## Custom`, and `## Notifications`
+- `tasks/deploys.md` has environment headings with dated deployment ledger entries including branch, commit range, commit count, and status
+- `.agents/project.json`, when present, is valid JSON and includes project designation fields used by pack-aware skills
+
+### 5. Audit Codex Mirror (`codex` scope)
+
+Compare Claude and Codex directories within each root:
+
+- Every skill in `global/claude/` should have a corresponding `global/codex/<name>/SKILL.md`
+- Every skill in `packs/<pack>/claude/` should have a corresponding `packs/<pack>/codex/<name>/SKILL.md`
+- Flag skills that exist in one but not the other
+- For skills that exist in both, check that the codex version has:
+  - Matching frontmatter `name` and `description`
+  - An `agents/openai.yaml` manifest (if the skill is execution-oriented, not display-only)
+
+**Known exceptions** (skills that intentionally only exist in one platform):
+- Check for a `# codex-skip` comment in the claude skill's frontmatter — if present, don't flag the missing codex mirror
+
+### 6. Generate Report
+
+Categorize all findings:
+
+| Severity | Meaning |
+|----------|---------|
+| **Error** | Convention violation that should be fixed (missing required field, naming mismatch) |
+| **Warning** | Drift or gap that may be intentional (missing codex mirror, incomplete sections) |
+| **Info** | Suggestions for improvement (long descriptions, missing optional sections) |
+
+For documentation templates, use:
+
+- **Error** for automation-breaking structure: missing checkboxes in `tasks/todo.md`, missing manual blocker annotations, executable work misfiled in `tasks/manual-todo.md`, `tasks/record-todo.md`, or `tasks/recurring-todo.md`, malformed phase numbering, multiple H1 headings, invalid `.agents/project.json`, or missing required task/spec sections that downstream skills parse.
+- **Warning** for template drift: missing metadata, missing `## Next Steps`, missing spec acceptance criteria, research docs without source/evidence orientation, or legacy roots that should move to canonical locations.
+- **Info** for uncertain classifications, old hand-written docs, optional sections, or cleanup suggestions that do not block automation.
+
+### 7. Auto-Fix (if `fix` mode)
+
+Only fix mechanical, unambiguous issues:
+
+- Add missing `argument-hint:` field (empty) to frontmatter
+- Fix `name` field to match directory name
+- Create missing Codex mirror in the matching root from the Claude version (copy frontmatter + intro, add TODO for Codex-specific content)
+- Add missing skill entries to the appropriate global core table or pack list in `docs/skills-reference.md`
+- Add missing empty metadata scaffold only when the target generated-doc template explicitly requires metadata and the values can be left blank
+- Normalize duplicate blank lines around headings in generated docs when it is purely mechanical
+
+**Never auto-fix:**
+- Missing Process/Output/Constraints sections (these need human judgment)
+- Content mismatches between claude and codex versions
+- Roadmap or todo content
+- Summaries, sources, acceptance criteria, research claims, history entries, task status, or next-step recommendations
+- Spec/research document rewrites that should use the archive-first replacement policy from the generating skill
+
+After fixing, re-run the audit to show remaining issues.
+
+## Output
+
+Display directly to the user (no files written in audit mode):
+
+```
+## Hygiene Report — [scope]
+
+### Errors (X)
+- **global/claude/foo/SKILL.md** — missing `version` field in frontmatter
+- **packs/business-discovery/claude/bar/SKILL.md** — directory name `bar` doesn't match frontmatter name `baz`
+
+### Warnings (X)
+- **packs/example-pack/codex/example-skill/** — missing Codex mirror (exists in Claude only)
+- **docs/skills-reference.md** — `$hygiene` not listed in the global core table
+- **research/icp.md** — missing `## Next Steps` section required by the research doc template
+
+### Info (X)
+- **global/claude/ship/SKILL.md** — no `## Constraints` section (consider adding)
+- **docs/architecture.md** — unknown hand-written doc; skipped strict generated-doc template checks
+
+### Summary
+- Skills: 48 checked, 2 errors, 1 warning
+- Tasks: 4 checked, 0 errors
+- Docs: 3 checked, 1 warning
+- Codex: 46 checked, 2 warnings
+```
+
+In `fix` mode, prepend each fixed item with a checkmark:
+
+```
+### Fixed
+- [x] Added missing `argument-hint:` to global/claude/foo/SKILL.md
+- [x] Created mirror global/codex/hygiene/SKILL.md
+
+### Remaining Errors (X)
+...
+```
+
+## Constraints
+
+- **Read-only by default.** Only modify files when explicitly invoked with `fix` mode.
+- **No content generation.** Auto-fix only adds structural scaffolding, never writes substantive content.
+- **Structural docs only.** Template checks validate shape, parseability, and canonical placement; use `$reconcile-dev-docs`, `$reconcile-research`, `$research-roadmap`, or `$spec-drift` for truth, freshness, and contradictions.
+- **Show evidence.** Every finding must include the specific file path and what's wrong.
+- **No false positives.** If uncertain whether something is a violation, classify it as Info, not Error.
+- **Respect exceptions.** Check for `# codex-skip` and similar markers before flagging intentional gaps.
+- **Use subagents** to parallelize scanning across skills, tasks, docs, and codex scopes when running `all`.
+
+
+## Default Shipping Contract
+
+Follow the shared shipping contract convention in CLAUDE.md.

package/packs/docs-health/codex/hygiene/references/documentation-templates.md

@@ -0,0 +1,248 @@
+# Documentation Template Contract
+
+Use this reference when `$hygiene` runs with `docs` or `all` scope. The goal is to check structure, parseability, and canonical placement for generated project documents. Do not use this audit to decide whether document claims are true, current, or complete.
+
+## Classification
+
+Classify files by path before checking headings:
+
+| Family | Paths | Strictness |
+| --- | --- | --- |
+| Task pipeline | `tasks/roadmap.md`, `tasks/todo.md`, `tasks/manual-todo.md`, `tasks/record-todo.md`, `tasks/recurring-todo.md`, `tasks/history.md`, `tasks/phases/phase-N.md`, `tasks/ideas.md`, `tasks/handoff.md`, `tasks/lessons.md` | Strict for workflow-owned files |
+| Specs | `specs/*.md`, `specs/{app}/*.md`, `docs/specifications/*.md` | Strict, except `*-interview.md` |
+| Interview logs | `*-interview.md`, `research/*-interview.md`, `research/{app}/*-interview.md` | Moderate |
+| Research docs | `research/*.md`, `research/{app}/*.md` | Strict for known skill outputs |
+| Research support logs | `research/*-search-log.md`, `research/{app}/*-search-log.md` | Moderate |
+| Experiments | `research/experiments/*.md`, `research/{app}/experiments/*.md` | Strict |
+| Audit/report docs | `specs/drift-report.md`, `tasks/reconciliation-report.md`, `research/reconciliation-report.md`, `docs/debug-changelog.md` | Moderate; append-only reports may vary |
+| Ops docs | `sync.md`, `deploy.md`, `tasks/deploy.md`, `tasks/deploys.md` | Strict where headings are named below |
+| Agent/project docs | `CLAUDE.md`, `AGENTS.md`, `.agents/project.json`, `README.md` | Structural only |
+| Skill-library docs | `SKILL.md`, `PACK.md`, `docs/skills-reference.md` | Use the existing skill and skills-reference checks first |
+| Alignment review pages | `alignment/*.html` | Generated browser-review artifacts; placement-only check |
+
+Skip strict template checks for `docs/history/archive/**`. Treat `alignment/*.html` as canonical generated HTML review artifacts, not Markdown documentation. Treat unknown hand-written Markdown as Info unless a path or heading clearly identifies it as a generated workflow document.
+
+## Universal Checks
+
+- First non-empty Markdown heading is exactly one `#` H1.
+- Generated files have at most one H1.
+- Required sections for the file family appear once.
+- Required sections appear in the expected order when downstream skills parse them.
+- Checkable workflow docs use `- [ ]` or `- [x]` for actionable items.
+- Non-ledger research/spec/task artifacts should include a metadata quote block when the family template requires it.
+- Main research/spec docs end with `## Next Steps`, or explicitly state why no next step exists.
+- Canonical roots are `tasks/`, `specs/`, `research/`, and generated HTML review pages under `alignment/`. `docs/specifications/` is a fallback spec location only.
+- New `docs/plan.md` or `docs/phases/` planning artifacts are legacy drift because current workflows write plans under `tasks/`.
+
+## Task Pipeline Templates
+
+### `tasks/roadmap.md`
+
+Required:
+- H1 title.
+- `## Summary` or equivalent orientation section.
+- Phase overview table or clearly labeled phase list.
+- Repeated `## Phase N:` sections.
+- Each phase has a milestone or acceptance criteria section.
+- Multi-phase roadmaps include `## Cross-Phase Concerns`.
+
+Flag as Error when phase numbering is malformed or downstream skills cannot identify phases. Flag as Warning when older roadmaps use equivalent but non-standard headings.
+
+### `tasks/todo.md`
+
+Required:
+- H1 title or current phase heading.
+- Checkable work items, `## Priority Task Queue`, or `## Priority Documentation Todo`.
+- Current-phase scope only, unless the file is visibly in legacy migration state.
+
+Flag as Error when there are no checkboxes and no priority queue. Flag as Warning when it appears to contain the full roadmap instead of the active phase.
+
+### `tasks/manual-todo.md`
+
+Required:
+- Checkable human-only external tasks.
+- Every unchecked item includes `_(blocks: Step N.X)_` or `_(after: Step N.X)_`.
+- No agent-executable repo/code/config/test/audit/CLI/API work.
+
+Missing blocker/after annotations are Errors because `$exec` and `$ship` use them to avoid executing blocked work. Agent-executable work in `tasks/manual-todo.md` is also an Error because it hides runnable work from the execution queue.
+
+### `tasks/record-todo.md`
+
+Required for each unchecked item:
+- Checkable one-time record or measurement task.
+- Source.
+- Condition that makes the record eligible.
+- Non-blocking reason.
+- Required data or access.
+- Measurement/query or exact evidence to collect.
+- Target or acceptance note.
+- Revisit cadence/date.
+- Completion evidence.
+- Promotion rule for when the item should move into `tasks/todo.md`.
+
+Flag as Error when the file contains executable implementation work that belongs in `tasks/todo.md`, or manual blockers that belong in `tasks/manual-todo.md`. Flag missing required fields as Warning unless downstream automation depends on them.
+
+### `tasks/recurring-todo.md`
+
+Required for each unchecked or active item:
+- Recurring task.
+- Cadence.
+- Owner/agent.
+- Scope.
+- Trigger.
+- Last run.
+- Next due.
+- Command/skill to run.
+- Evidence/output path.
+- Escalation conditions.
+
+Flag as Error when recurring items are written as current execution steps without an explicit promotion rule. Flag missing cadence, next due, or evidence path as Warning.
+
+### `tasks/history.md`
+
+Required:
+- H1 title.
+- Append-only dated entries, preferably `## YYYY-MM-DD - summary`.
+
+Do not require old entries to be rewritten. Missing dates are Warnings unless they prevent reconciliation.
+
+### `tasks/phases/phase-N.md`
+
+Required:
+- H1 phase title.
+- Completed steps.
+- Milestone or acceptance criteria.
+- `## On Completion` or equivalent completion summary.
+
+## Spec Templates
+
+### Implementation specs
+
+Applies to `specs/*.md`, `specs/{app}/*.md`, and fallback `docs/specifications/*.md`, excluding interview logs and generated audit reports.
+
+Required sections:
+- `## Overview`
+- `## Goals`
+- `## Non-Goals`
+- `## Detailed Design`
+- `## Edge Cases`
+- `## Test Plan`
+- `## Acceptance Criteria`
+- `## Open Questions` or explicit `None`
+
+Equivalent headings may be Warning instead of Error if the document is older but still readable. Missing acceptance criteria or test plan is Warning unless another workflow depends on it directly.
+
+### Interview logs
+
+Required:
+- H1 `Interview Log: <topic>` or equivalent.
+- Questions asked.
+- Options presented when options were used.
+- User responses or decisions.
+- Final deviations from the initial draft or coverage summary.
+
+## Research Templates
+
+### Main research docs
+
+Applies to known research-skill outputs such as `research/icp.md`, `research/competitive-analysis.md`, `research/journey-map.md`, `research/metrics.md`, `research/gtm.md`, `research/monetization.md`, `research/customer-feedback.md`, `research/enterprise-icp.md`, and app-scoped equivalents.
+
+Required:
+- H1 title.
+- Metadata quote block with `Last updated`, `Status`, and `Sources` when practical.
+- `## Summary`.
+- Body sections matching the producing skill's documented output.
+- Source/evidence orientation for factual claims.
+- Assumptions, risks, or validation gaps when relevant.
+- Final `## Next Steps`.
+
+Missing `Next Steps`, metadata, or source orientation is Warning by default.
+
+### Search logs
+
+Required:
+- H1 title identifying the log.
+- Query list or table.
+- Findings.
+- Source attribution.
+- Scoring or reasoning notes when the producing skill requires them.
+
+Search logs are supporting context, not primary completion artifacts.
+
+### Experiments
+
+Required:
+- `## Hypothesis`
+- `## Method`
+- `## Success Criteria`
+- `## Timeline`
+- `## Budget`
+- `## Decision Rules`
+- `## Results`
+- `## Next Steps`
+
+## Audit And Ops Templates
+
+### Reconciliation and drift reports
+
+Applies to `tasks/reconciliation-report.md`, `research/reconciliation-report.md`, and `specs/drift-report.md`.
+
+Required:
+- H1 report title.
+- Resolved findings.
+- Deferred findings.
+- Remaining findings or confirmation that none remain.
+- Evidence references.
+
+### `docs/debug-changelog.md`
+
+Required per entry:
+- Date.
+- Symptom.
+- Category and severity.
+- Root cause.
+- Fix.
+- Test results.
+- Related entries.
+- Systemic flag.
+
+### `sync.md`
+
+Required top-level sections:
+- `## Dependencies`
+- `## Conflict Resolution`
+- `## Custom`
+- `## Notifications`
+
+Shell commands should be fenced code blocks under executable sections.
+
+### `tasks/deploys.md`
+
+Required:
+- Environment headings.
+- Dated deployment entries.
+- Branch.
+- Commit range.
+- Commit count.
+- Status: success or failed.
+
+Failed deploys may be recorded but should not reset staleness.
+
+## Agent And Project Docs
+
+- `CLAUDE.md` must exist.
+- If `AGENTS.md` exists, it should not contradict `CLAUDE.md` on workflow orchestration, task roots, or shipping policy.
+- `.agents/project.json`, when present, must be valid JSON and include project designation fields used by pack-aware skills.
+- `README.md` should have overview, installation or usage, structure, and validation sections, but missing sections are Info unless this repository's conventions make them required.
+
+## Severity Guide
+
+- Error: automation-breaking structure, invalid JSON in machine-read files, missing task state in workflow-owned docs, malformed phase numbering, or missing blocker annotations.
+- Warning: template drift, missing metadata, missing `Next Steps`, missing acceptance criteria, missing source orientation, legacy roots, or older equivalent headings.
+- Info: unknown hand-written docs, optional cleanup, uncertain classification, or style improvements.
+
+## Fix Policy
+
+Auto-fix only mechanical structure. It is safe to add empty metadata scaffolding when the template explicitly requires metadata and values can be left blank. It is safe to normalize duplicate blank lines around headings.
+
+Never auto-generate summaries, sources, acceptance criteria, research claims, history entries, task status, or next-step recommendations. Never rewrite specs or research docs to fit a template; report those findings and let the generating/reconciliation skill handle contentful changes.