skills/pickup-repo/SKILL.md
--- name: pickup-repo description: Repo-side session re-entry, the back half of /ro:graduate-project. On a fresh session in a code repo it reads the repo CLAUDE.md then docs/agents/HANDOFF.md, refreshes live state from `gh issue list --label ready-for-agent` plus the coordination-hub issue, follows the cross-vault back-pointer to read the gestation vault's recent ROADMAP / log / decisions, and prints a one-screen briefing (what this is, where it stopped, what's next, what's blocked, what to clai
npx skillsauth add RonanCodes/ronan-skills skills/pickup-repoInstall 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.
/ro:graduate-project is the forward half of the LLM-Wiki to repo seam: when
a project leaves ideation it writes docs/agents/HANDOFF.md, opens a
coordination-hub GitHub issue plus sliced ready-for-agent child issues, and
writes a thin repo CLAUDE.md whose "Start here" block points at both with a
cross-vault back-pointer to the gestation vault.
This is the back half: the deliberate repo-side re-entry. It is the
project-side equivalent of the vault [[pickup]]. A fresh session in a graduated
repo can self-orient from the generated CLAUDE.md "Start here" block (that
covers most of it), but this skill makes the re-entry a single command that also
does the vault round-trip: it reads the gestation vault's recent state so the
briefing reflects what happened on the wiki side since graduation, not just the
snapshot frozen into HANDOFF.md.
<repo> llm-wiki-<slug> vault
CLAUDE.md "Start here" ──pickup-repo──▶ ROADMAP / log / decisions
docs/agents/HANDOFF.md (read recent state)
GH hub issue + ready-for-agent issues
│
└──── one-screen briefing ◀──── reconcile all three sources
/ro:pickup-repo./ro:graduate-project has run on a repo and a later session needs to
start from cold.Do NOT fire for:
HANDOFF.md, no hub issue). Fall back to
reading CLAUDE.md + README.md + gh issue list and say the repo was not
graduated, there is no vault round-trip to do.[[pickup]]. This is the repo side.gh repo view --json nameWithOwner if a gh remote
exists). --repo <owner/name> overrides.docs/agents/HANDOFF.md.
If it is missing, degrade gracefully (see "Not graduated" below).In order, exactly as the generated "Start here" block intends:
CLAUDE.md: the thin repo guide (one-liner, stack, live URL,
load-bearing gotchas, ship flow, "Next up", and the cross-vault back-pointer).
Pull the back-pointer (the obsidian://open?... link or the named vault). It
is what step 4 needs.docs/agents/HANDOFF.md: the cold-start doc (what it is, architecture
north-star, decisions made AND rejected, open questions, blockers, backlog
pointers, conventions). This is the frozen snapshot from graduation, treat it
as the baseline the live sources update.The handoff is a snapshot; the issues are the truth.
gh issue list --label ready-for-agent --state open --json number,title,labels
HANDOFF.md / CLAUDE.md, the
#N-style shared-memory anchor). It states the work split and the
shared-memory model.ready-for-agent issues first; note anything labelled
blocked or blocked-on-human and why.--no-vault)This is the round-trip that makes this skill more than the "Start here" block.
bash "${CLAUDE_PLUGIN_ROOT:-$HOME/Dev/ronan-skills}/scripts/resolve-repo.sh" LLM_WIKI)
and the project's gestation vault from the back-pointer in CLAUDE.md (or
./.ronan-skills-llm-wiki.local.json).ROADMAP.md (In progress / Next up / Blocked /
recently done), the tail of log.md (decisions since graduation), and any
decisions / open-questions pages that changed after the graduation date./query "<project> recent decisions + open questions + roadmap" --vault llm-wiki-<slug> over raw grep so the progressive index surfaces the
load-bearing pages.HANDOFF.md predates, or a
ROADMAP item that has no matching ready-for-agent issue yet.Reconcile all three sources into a single screen a fresh agent reads cold:
PICKUP: <repo> (graduated from llm-wiki-<slug>)
What it is: <one line from HANDOFF / CLAUDE>
Where it is: <live URL / stack / current branch>
Stopped at: <last thing done, from HANDOFF + recent vault log>
In flight: <open issues / branches mid-work>
Next up: <top ready-for-agent issues: #.. #..>
Blocked: <blocked-on-human / paid-resource / external party + why>
To claim: <the issue(s) a worker should pull first>
Drift: <vault decisions newer than HANDOFF, if any>
Keep it to one screen. Detail lives in the linked HANDOFF, hub issue, and vault , the briefing is the map, not the territory.
Use AskUserQuestion to ask which issue(s) to start, then hand off to the right
flow (/ro:ralph, /ro:planner-worker, or just claim a single issue). The heavy
reading happened in this pickup; the main context stays a thin router.
If there is no HANDOFF.md / hub issue, this repo was not graduated. Don't bail:
read CLAUDE.md + README.md + gh issue list --state open, print the shortened
briefing (no vault round-trip), and tell the user the repo was never graduated ,
suggest /ro:graduate-project if it gestated in a vault.
--repo <owner/name>: target repo if not the current directory.--no-vault: skip the cross-vault round-trip (repo + GitHub only).--brief-only: print the briefing and stop; do not offer the next move.[[graduate-project]]: the forward half this completes. That writes the
orientation layer; this re-enters through it with the vault round-trip.[[pickup]]: the vault-side session-start briefing. This is the repo-side
equivalent.[[handoff]] (handoff resume): the per-session repo re-entry from a
.handoff/ file. pickup-repo is the cross-phase, graduated-project version
that always does the vault round-trip.[[close-session]]: the durability checklist that names this skill as the
on-the-roadmap repo-side pickup.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".