hir4ta
UserQuality Enforcement Layer for Claude Code — Hook-gated SDD (spec & plan vehicles), Wave commits, and grounded multi-agent code review.
Categories
Indexed Skills (17)
reflect
Generate a retrospective markdown for an archived (or about-to-be-archived) mumei feature. Reads requirements / design / tasks / spec-reviews / reviews / cost-log and emits reflect.md with AC count, Wave count, review iter counts, fix-spiral detection, token cost, cache hit rate, and hook firing breakdown. Triggered by the user invoking /mumei:reflect <feature> after /mumei:retire (or before, if explicitly requested). Read-only with respect to feature content; writes only the reflect.md output file.
retire
Moves a completed feature directory to .mumei/archive/<YYYY-MM>/<feature>/ once the feature reaches phase=done. Auto-detects the active vehicle by checking .mumei/specs/<feature>/ first, then .mumei/plans/<feature>/. Triggers when the user explicitly retires a feature or when /mumei:proceed or /mumei:examine finishes with verdict=PASS and the user confirms.
muse
Generate a retrospective markdown for an archived (or about-to-be-archived) mumei feature. Reads requirements / design / tasks / spec-reviews / reviews / cost-log and emits muse.md with AC count, Wave count, review iter counts, fix-spiral detection, token cost, cache hit rate, and hook firing breakdown. Triggered by the user invoking /mumei:muse <feature> after /mumei:shelve (or before, if explicitly requested). Read-only with respect to feature content; writes only the muse.md output file.
review
Use when the user wants a one-shot, strongest-available code review of the current diff WITHOUT the full mumei spec/plan workflow — including users who drive their own spec or are doing a change too small to plan. Reviews `git diff $(git merge-base <base> HEAD)` (PR-pushed commits + uncommitted changes) through the shared engine: deterministic detectors (semgrep / osv / secret / type-check / test) → diverse LLM reviewers (security + adversarial, plus spec-compliance when a spec file is passed) → single per-finding adjudication gate → class-aware fail-open verdict. Has ZERO side effects: it never creates `.mumei` state, never writes the ledger or agent memory, and never commits. Triggers on "/mumei:review", "review my diff", "review this PR locally". For the full SDD workflow use /mumei:compose; for plan-vehicle review use /mumei:peruse.
shelve
Moves a completed feature directory to .mumei/archive/<YYYY-MM>/<feature>/ once the feature reaches phase=done. Auto-detects the active vehicle by checking .mumei/specs/<feature>/ first, then .mumei/plans/<feature>/. Triggers when the user explicitly shelves a feature or when /mumei:compose or /mumei:peruse finishes with verdict=PASS and the user confirms.
assure
Render a detailed reliability view of a mumei feature — pass^3 over the most recent 10 trials plus a table of the last 10 trial rows from reliability-log.jsonl. Triggered only by explicit user invocation `/mumei:assure <feature>`. Reads .mumei/specs/<feature>/reliability-log.jsonl or .mumei/plans/<feature>/reliability-log.jsonl via hooks/_lib/reliability.sh. Exits non-zero with `feature not found: <feature>` to stderr when the feature directory is absent. The k=3 / window=10 parameters are fixed.
present
Render a one-line reliability summary of the active mumei feature (or a specified one). Prints `<feature> | pass^3: <value-or-N/A> (n=<n_trials>, window=10, k=3)` to stdout — exactly one line, no headers. Triggered only by explicit user invocation `/mumei:present` (no args reads `.mumei/current`) or `/mumei:present <feature>`. Outputs `no active feature` to stdout (not stderr) and exits 0 when `.mumei/current` is missing or points to a non-existent feature. Read-only via hooks/_lib/reliability.sh.
attest
Render a detailed reliability view of a mumei feature — pass^3 over the most recent 10 trials plus a table of the last 10 trial rows from reliability-log.jsonl. Triggered only by explicit user invocation `/mumei:attest <feature>`. Reads .mumei/specs/<feature>/reliability-log.jsonl or .mumei/plans/<feature>/reliability-log.jsonl via hooks/_lib/reliability.sh. Exits non-zero with `feature not found: <feature>` to stderr when the feature directory is absent. The k=3 / window=10 parameters are fixed.
glance
Render a one-line reliability summary of the active mumei feature (or a specified one). Prints `<feature> | pass^3: <value-or-N/A> (n=<n_trials>, window=10, k=3)` to stdout — exactly one line, no headers. Triggered only by explicit user invocation `/mumei:glance` (no args reads `.mumei/current`) or `/mumei:glance <feature>`. Outputs `no active feature` to stdout (not stderr) and exits 0 when `.mumei/current` is missing or points to a non-existent feature. Read-only via hooks/_lib/reliability.sh.
arrange
One-time setup for a project to use mumei. Detects existing CLAUDE.md / .claude/rules/, proposes additions about mumei's expectations (phase gates, Wave commits, review pipeline), and applies them with user approval. Triggers when the user says "set up mumei", "install mumei", or "initialize mumei in this project".
examine
Plan-vehicle review pipeline. Runs Stage 0 detector (semgrep + osv-scanner) plus security-reviewer and adversarial-reviewer in parallel against the current diff, validates each finding via issue-validator, aggregates a verdict, and writes a review JSON to `.mumei/plans/<slug>/reviews/<ts>.json`. Triggers when the user invokes /mumei:examine on a plan-vehicle feature whose state.json has `pending_review=true` (set automatically when the last TaskCompleted matches task_created_count). On PASS, advances `phase` to `done` and prompts the user to run /mumei:retire. On MAJOR_ISSUES, surfaces findings and leaves session-end and git-push blocks active until they are addressed.
gather
This skill should be used BEFORE any feature design. It runs structured gathering with the user — asking 5 high-leverage questions per round, up to 3 rounds, to extract Goal / Scope / Constraints / Edges / Done. Output is saved to .mumei/scratch/<topic>.md and used as input for /mumei:proceed. Triggers include "I want to add X", "we need a feature for Y", "let's brainstorm Z", or any vague feature request.
proceed
The mumei orchestrator. For new features, presents a vehicle picker — `spec` (full SDD workflow: clarification → requirements → design → tasks each auto-reviewed up to 3 iterations → single user approval → Wave-by-Wave implementation → 4-stage review) or `plan` (Claude Code plan-mode wrapper: hand off to plan mode, capture via hook, run /mumei:examine at the end). Resumes existing features automatically by detecting which vehicle's state.json exists. Triggers when the user invokes /mumei:proceed <feature> or naturally asks to "plan", "spec", "design", or "implement" a feature with mumei. Always renders body content (User Story prose, AC bodies after EARS keywords, Assumptions, Open Questions, design narratives, task descriptions, Wave goals/verifies) in the user's conversation language; English section headings, EARS keywords, REQ trace IDs, and [CONFIRMED]/[ASSUMPTION] annotations remain literal.
compose
The mumei orchestrator. For new features, presents a vehicle picker — `spec` (full SDD workflow: clarification → requirements → design → tasks each auto-reviewed up to 3 iterations → single user approval → Wave-by-Wave implementation → 4-stage review) or `plan` (Claude Code plan-mode wrapper: hand off to plan mode, capture via hook, run /mumei:peruse at the end). Resumes existing features automatically by detecting which vehicle's state.json exists. Triggers when the user invokes /mumei:compose <feature> or naturally asks to "plan", "spec", "design", or "implement" a feature with mumei. Always renders body content (User Story prose, AC bodies after EARS keywords, Assumptions, Open Questions, design narratives, task descriptions, Wave goals/verifies) in the user's conversation language; English section headings, EARS keywords, REQ trace IDs, and [CONFIRMED]/[ASSUMPTION] annotations remain literal.
glean
This skill should be used BEFORE any feature design. It runs structured gleaning with the user — asking 5 high-leverage questions per round, up to 3 rounds, to extract Goal / Scope / Constraints / Edges / Done. Output is saved to .mumei/scratch/<topic>.md and used as input for /mumei:compose. Triggers include "I want to add X", "we need a feature for Y", "let's brainstorm Z", or any vague feature request.
kindle
One-time setup for a project to use mumei. Detects existing CLAUDE.md / .claude/rules/, proposes additions about mumei's expectations (phase gates, Wave commits, review pipeline), and applies them with user approval. Triggers when the user says "set up mumei", "install mumei", or "initialize mumei in this project".
peruse
Plan-vehicle review pipeline. Runs Stage 0 detector (semgrep + osv-scanner) plus security-reviewer and adversarial-reviewer in parallel against the current diff, validates each finding via issue-validator, aggregates a verdict, and writes a review JSON to `.mumei/plans/<slug>/reviews/<ts>.json`. Triggers when the user invokes /mumei:peruse on a plan-vehicle feature whose state.json has `pending_review=true` (set automatically when the last TaskCompleted matches task_created_count). On PASS, advances `phase` to `done` and prompts the user to run /mumei:shelve. On MAJOR_ISSUES, surfaces findings and leaves session-end and git-push blocks active until they are addressed.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.