← ClaudeAtlas

73listed

Sign-off — end-of-session checklist. Confirms it's safe to end, surfaces open action items, and writes the session's learnings back to memory and the KB so the workspace grows. Triggers on '/73', 'signing off', 'wrap up the session', 'end of session', 'that's a wrap'.
liormesh/trestle · ★ 2 · AI & Automation · score 73
Install: claude install-skill liormesh/trestle
# /73 — Sign-off The user is signing off (`/73` — ham radio for "best regards / signing off"). Run a three-part end-of-session check and reply with the results. This is the ritual that makes the workspace **grow**. "Corrections become memory, finished work accumulates in the knowledge base" only happens if something writes it back at the end of a session — this is that something. Don't skip the writing. ## The check 1. **Safe to end?** — Anything mid-flight: uncommitted edits, running processes, half-finished tool calls, undeployed changes? 2. **Open action items?** — What's left for the user, or for you to follow up on next session? 3. **Documented everything?** — Memory updated, KB notes written, task-tracker cards created/updated, any other persistence the session warrants? **Don't just check — actually do the documentation.** Scan the thread for durable value (decisions, learnings, project-state changes, new context, surprising findings) and write each to its correct home *before* you answer: - **Cross-cutting behavioral correction** → inline in MEMORY.md, or a `claude-memory/feedback_*.md` only if it genuinely cross-cuts most sessions. - **Project state / decision** → `projects/<project>/overview.md`. - **Domain learning** → the relevant book chapter. - **Follow-up task** → your PM tool (Trello / Linear / Jira / GitHub Issues / whatever you configured). Follow the two standing rules while you write: **no standalone feedback files** (inline it next t