end-task
Skill7 starsChanged 49 days agoMITView on GitHub
Lightweight per-chat close-out check, distinct from end-the-day's fuller once-daily sweep. Checks whether Claude's own auto-memory files relevant to this session's work are still accurate - updates anything stale (predates a change made this session) or duplicated (an outstanding item now resolved), and reports what changed. Does not touch vault notes, the task manager, or brag items - those stay end-the-day's job. Trigger: /end-task, or saying you're about to close this chat and want to check notes are up to date.
In this repository12 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- end-the-daySkill
- link-taskSkill
- log-completed-tasksSkill
- pick-meetings-to-prepSkill
- prep-meetingSkill
- search-mailSkill
- setupSkill
- start-the-daySkill
- suggest-brag-itemsSkill
- surface-stale-tasksSkill
- update-tasksSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 76
No em-dashes in anything written - hyphens instead (edit to taste).
- Distinctive: line 77
Prefer bullet-point lists over walls of prose.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
End Task
Purpose
A pattern that works well with these skills: keep one long-running "daily"
chat for ad hoc questions (see start-the-day), and fork each substantial
task into its own chat, closing it once the task's done. Before closing a
task chat, it's worth making sure Claude's own auto-memory - its persistent
memory files, not the vault - actually reflects what just happened, so the
next chat you open starts from accurate context instead of stale carry-over.
This is deliberately narrower than end-the-day. That skill does a heavier
vault-wide sweep (loose ends across every note edited today, task-manager
cross-checks, brag-item suggestions) that's worth the cost once a day, not
every time a task chat closes. end-task only looks at memory freshness -
cheap enough to run on every close, and the thing that most directly
protects the next chat.
Trigger
/end-task, or saying you're closing this chat and want to confirm
notes/memory are current before you do.
Process
- Work out what this session actually touched - which project(s) or topic(s) the conversation's work relates to. No vault-wide scan needed; this is scoped to the current session's content.
- Check the relevant existing memory files for that project/topic (via
the memory index, e.g.
MEMORY.md, and the files it points to). Look specifically for:- Anything stated as "not yet done," "outstanding," or "blocked" that this session actually resolved.
- Anything this session's new information supersedes or contradicts.
- Duplicate entries that should collapse into one once updated.
- Cross-check against anything already tagged in today's Daily note for the session's topic (e.g. a digest tag you use for weekly project updates), purely to confirm memory and the note agree on what happened - don't add or edit tags here, that's a vault edit and out of scope.
- Update the affected memory files directly. Memory writes don't need approval the way vault edits do - but say plainly what was written, corrected, or removed, so the user can see the diff rather than just being told it's handled.
- Report a short close-out summary: what was checked, what (if anything) changed, and confirmation it's safe to close. If nothing needed updating, say so briefly rather than padding it out.
Out of scope
- Vault notes (Daily note, project notes) - loose ends and wording there are
end-the-day's job. - The task manager - no reads or writes here at all.
- Brag-list suggestions - stays
suggest-brag-items's job, run fromend-the-day. - Adding or editing digest tags - this skill only checks memory against what's already tagged.
Notes on conventions
- No em-dashes in anything written - hyphens instead (edit to taste).
- Prefer bullet-point lists over walls of prose.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

