← ClaudeAtlas

project-handofflisted

Use when the user asks to update or prepare the handoff, wrap up a session, or switch chats, or when the context is getting large. Writes findings to the repo docs, updates docs/HANDOFF.md, and commits.
Trooper-code/Trooper-s-Coordinator-Skills · ★ 2 · AI & Automation · score 75
Install: claude install-skill Trooper-code/Trooper-s-Coordinator-Skills
# Project handoff (end of session / chat switch) Goal: nothing learned in this session is lost, and the paste into the next chat stays small. The next chat starts from `docs/HANDOFF.md` (plus the coding agent's latest summary, if you use one) and must know exactly what to do first. Principle: **the repo docs are the record; HANDOFF.md is a short index into them.** Every new chat pays for every line of HANDOFF.md, so it holds the current state and the next step, not history. Budget: about 100 lines / 10 KB. ## Files this skill expects (create them if missing) - `docs/HANDOFF.md` — the handoff itself. - `docs/MILESTONES.md` — the status board: milestones, a dated progress log, and the "Next" step. - Research docs — `RESEARCH.md` or `docs/research/*.md` — where findings live (formats, offsets, protocols, how things work). - `docs/prompts/` — prompts for the coding agent (see the agent-coordinator skill). - `docs/reports/` — the coding agent's reports (it commits them on its branch), with `TEMPLATE.md` for the format. - `AGENTS.md` — the coding agent's standing rules (see the agent-coordinator skill). - `docs/DECISIONS.md` — real decisions: what was chosen, what was rejected, why. - `docs/research/INDEX.md` — one line per research topic, pointing to its file and section. - `scripts/check-handoff.py` — checks HANDOFF.md's size, sections, ending and paths, and notes lessons to promote. - `scripts/check-research.py` — lists confirmed facts proven on older tool versions, and broke