skills/stay-in-your-lane/SKILL.md
--- name: stay-in-your-lane description: Scope each dev's or agent's work on a shared repo so parallel workers don't collide. Reads the source of truth (Linear / GitHub issues / Jira / Trello) or asks who is on what, then maintains a committed LANES.md in the target repo that maps each person or agent to the files, directories, and tasks they own. Auto-applies whenever LANES.md is present - before touching a file, an agent checks whether it falls in another lane and either steps around it or ask
npx skillsauth add RonanCodes/ronan-skills skills/stay-in-your-laneInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
When several people or agents (all on Claude) work one repo at once, the failure mode is two
of them editing the same area and clobbering each other. This skill draws lanes: a committed
LANES.md that says who owns which files, directories, and tasks. Before an agent touches a
file, it checks the lanes and either steps around someone else's area or asks first.
It's the soft, human-readable counterpart to git worktrees. Worktrees isolate the working copy; lanes isolate the intent, so even agents sharing one checkout know where the edges are.
/ro:planner-worker, /ro:night-shift) where workers may overlap on
file areas and you want an explicit ownership map on top of the conflict detector.When a repo has a LANES.md at its root, this skill is in effect for the session without
being asked. At the start of work in such a repo:
LANES.md.--mine <name> arg, the ME marker in LANES.md, or
ask once and remember for the session).Before editing any file, resolve it against the lanes (see "Checking a path" below). If the file falls in another active lane, do not edit it silently - step around it, or surface the overlap to the user and ask.
A marker comment near the top of the target repo's CLAUDE.md / AGENTS.md makes this
discoverable: <!-- uses ronan-skills:stay-in-your-lane - see LANES.md -->. Treat
LANES.md itself as the real signal; the marker is a hint for humans skimming the repo.
--sync (or first run with no LANES.md) builds or refreshes the lane map. Resolve the source
in this order, or honour --source:
Map each owner to concrete paths (globs) plus a short task description, then write LANES.md.
Re-run --sync at the start of each sprint, or whenever someone changes focus, so the map tracks
reality rather than going stale.
A committed Markdown file at the repo root. Human-readable first, machine-parseable second. One section per lane.
# Lanes
> Ownership map for parallel work on this repo. Synced from Linear on 2026-06-28.
> Before editing a file, check it's in your lane. If it's in someone else's, ask first.
> Maintained by ronan-skills:stay-in-your-lane. Re-run `--sync` each sprint.
## Nicolas <!-- id: nicolas -->
- **Owns:** cart, product matching
- **Paths:**
- `src/cart/**`
- `src/matching/**`
- `src/lib/pricing.ts`
- **Tasks:**
- Cart persistence (LIN-204)
- Matching algorithm v2 (LIN-211)
- **Source:** linear:LIN-204, linear:LIN-211
## Ronan <!-- id: ronan, me -->
- **Owns:** everything else
- **Paths:**
- `src/**` (except Nicolas's paths above)
- `infra/**`
- **Tasks:**
- Auth + onboarding (LIN-198)
- **Source:** linear:LIN-198
## Shared <!-- id: shared -->
- **Owns:** files anyone may touch with care
- **Paths:**
- `package.json`
- `README.md`
- `.github/**`
- **Note:** coordinate before structural changes; small additive edits are fine.
Field reference per lane:
| Field | Meaning |
|---------|---------|
| heading | The owner's display name. The HTML comment carries the stable id and the me marker for the current user's own lane. |
| Owns | Plain-English summary of the area, for humans skimming. |
| Paths | Glob patterns this owner is responsible for. The load-bearing part - this is what path checks resolve against. |
| Tasks | Current tasks, ideally with tracker ids, so the map is auditable against the source of truth. |
| Source | Where the lane came from (linear:ID, github:#N, jira:KEY, trello:card, or ask). |
Resolution rules:
src/lib/pricing.ts in Nicolas's lane overrides src/** in Ronan's.(except ...) note on a glob is documentation; treat the explicit narrower lane as the real
owner when paths overlap.Shared lane is touchable by anyone, with the coordinate-first note honoured.--check <path> resolves a single file against the current LANES.md and reports its owner and
whether it's yours. Use this before any edit when working in a laned repo:
Shared or no lane: proceed (coordinate on Shared).For a swarm, each worker gets --mine <its-name> so it knows its own lane, and the planner
keeps LANES.md as the shared contract on top of the file-area conflict detector.
When focus shifts (someone finishes a lane, picks up a new area), update LANES.md and commit it
so the next agent or dev sees the change. Keep the Synced from ... on <date> line current. The
file is committed precisely so the ownership map travels with the repo and survives context resets.
LANES.md.LANES.md is the human-declared ownership layer that sits above it, so a dev
who never reads the swarm internals can still see and edit who owns what.testing
--- name: linear-pipeline description: The Fable orchestrator for a single dispatched Linear ticket. Holds almost no context itself; it receives `--issue <ID> --detached`, decides the stage sequence, and fans out a sub-agent per stage, passing forward only each stage's artifact (never re-derived, never inlined into its own context). Step zero, before any planning or stage routing, is a boundary triage against `canon/security-boundary.md` (#199): a match tags Ronan Connolly and stops the run, no
development
--- name: in-your-face description: Capture a chat-only answer into a durable artifact (markdown + HTML, PDF when cheap) and launch it automatically so the user cannot miss it. Use when user says "in your face", "don't let me lose this", "save that answer", "make that durable", or right after answering a substantive side question (a recipe, comparison, how-to, or generated prompt) that would otherwise die with the context. category: workflow argument-hint: [--no-open] [--vault <short>] [hint of
tools
One-shot headless OpenAI Codex CLI calls for background/admin AI tasks — summaries, classification, extraction, admin glue. The default engine for anything that runs AI constantly in the background (daemon-driven, per-event), because it bills the flat ChatGPT subscription instead of Claude usage or per-token API spend, and it keeps working while Claude is rate-limited. NEVER for coding — coding stays Claude. Use when a skill or daemon needs a cheap always-on AI call, when the user says "use codex", "ask codex", "codex as backup", or when building a background summarizer/classifier into a listener or loop. Reads auth from ~/.codex/auth.json (ChatGPT account, no API key).
research
Turn a warranty rejection, repair quote, or RMA email into a cited decision brief — legal read (NL/EU consumer law), is the part user-serviceable, live part and new-unit prices, repair-vs-DIY-vs-new economics, before-you-send-it checklist, deadlines. Use when the user pastes or screenshots a repair quote, warranty rejection, "not covered" email, onderzoekskosten fee, or asks "should I repair or replace this".