a-canary
UserUniversal agent harness — SQLite ledger + ephemeral interactive Claude Code workers
Categories
Indexed Skills (17)
analyse-recent-sessions
Read N recent worker tmux scrollbacks / handoff events, identify recurring friction, write a new skill or update an existing one.
bookie
Sole writer of ledger rows. Validates kind/type/state, mints slug ids, emits 'created' event. Adds `decompose` verb that turns a PRD into a chain of TDD slices.
clarify-docs
Hygiene phase skill. Fix doc/code drift, broken refs, stale ADRs, missing glossary terms, failed example commands. Docs-only PR, ≤3 files, ≤100 lines diff. Run via bookie create --kind task --class hygiene --title 'clarify-docs: <observation>'.
claude-afk
Headless-shaped invocation of `claude` that runs inside an interactive tmux pane you can attach to live. Gives you the shape of `claude -p --output-format json` with full transparency: every turn is observable in real time, attachable mid-flight, and inspectable after the fact via the session log.
deploy-preview
Cron-scheduled probe. Scans open issues with non-null pr_url, finds deploy preview URLs (Vercel/Netlify/Cloudflare Pages/GitHub Pages) in PR body+comments, emits deploy_preview events on the ledger so dev-quest's FOCUS pane can show a 'Preview ready' badge.
diff-review
Pre-commit phase. Independent subagent reviews the finalized diff against the task brief + touched ADRs, returns a structured report of consequences, surprises, gaps, and ADR conflicts. Worker logs the report as a ledger event before bookie will accept a merged state update.
improve-architecture
Slice-bounded refactor: clarify boundaries, remove incidental complexity, extract duplication. No behavior change.
prd-to-issues
Decompose a PRD (kind=prd row or PRD-*.md file) into N sprint ledger rows, each a tracer-bullet thin vertical slice with Requirements + Success criteria in body_md.
replay-shadow
Capture a real execution of a worker-shaped system, replay it against a candidate config in an isolated sandbox, diff the two. Use before promoting prompt changes, model swaps, skill set changes, or API upgrades. System-agnostic — works for any unit of work that has identifiable inputs, environment state, an observable transcript, and a measurable output.
spawn
Worker decomposition verb. Inserts one or more child ledger rows under the current task, wires parent.blocked_by to the new children, flips parent to blocked. Delegates the write to bookie. NOT a process spawn.
sprint-supervise
Re-entrant sprint supervisor loop. Drives a single thin vertical slice to evidence-backed done across re-entry cycles: load prior handoffs, check evidence, complete or decompose, write re-entry handoff, tear down.
trash-retired-files
Move files unused for ≥30d (or unreferenced per grep) into ~/trash/<unix-ts>_<name>-<YYYYMMDD>/ with a TTL marker. Reversible; the actual sweep is run by Slice C machinery.
triage-assign
Agent-actor triage loop. Batch-claims rows with agent='agent_unset' OR pool='pool_unset', assigns {tier, pool, agent} per payload + context, writes via bookie.
triage-failed
Director-subagent skill. Classifies a `state=failed` ledger row as low-risk (auto-decompose into slices, cancel parent) vs needs-HITL (mark for human review).
ke-learn
Write a new knowledge entry to ~/vault/ke/<scope>/ and update the FTS5 index. One file per insight. Stop-hook on dev/director sessions.
ke-recall
FTS5 search over ~/vault/ke/ knowledge entries. Returns top-N markdown excerpts ranked by BM25, with file paths and headings.
to-ledger
Owner-facing flow for filing a single ledger row. Walks the user through the required flags, then delegates the write to bookie.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.