close-stage
Skill1 starsChanged 4 months agoMITView on GitHub
Close a stage or numbered milestone (2.x, 3.x) in the researcher-harness — verify exit criteria, tick the ROADMAP, rewrite the status line, sweep problem.md, commit in the project's docs style. Use when exit criteria pass or a numbered feature ships.
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- commit-progressSkill
- debug-liveSkill
- evaluate-modelSkill
- exploreSkill
- implement-featureSkill
- run-testsSkill
- run-testsSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is 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.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Prove it first: Run whatever the exit criteria in docs/ROADMAP.md demand — uv run pytest -q at minimum.
- Distinctive: line 13
Tick the ROADMAP: Check the boxes with date + evidence ("7/7 gemini, 1,083 out-tokens") in the same style as the Stage 0–2 entries.
- Distinctive: line 16
Refresh the root
README.md: if the stage changed what the harness does or its numbers — its Status section hard-codes test counts and …
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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
Close a stage / milestone — the bookkeeping ritual
The ritual that keeps a fresh session able to orient from one status line. It has run twice already (docs: Stage 2 wrap, docs: Stage 2 CLOSED in git log) — same shape every time.
Steps
- Prove it first. Run whatever the exit criteria in
docs/ROADMAP.mddemand —uv run pytest -qat minimum; the live gate on the required provider(s) if the stage says so. No green, no close. - Tick the ROADMAP. Check the boxes with date + evidence ("7/7 gemini, 1,083 out-tokens") in the same style as the Stage 0–2 entries.
- Rewrite the status line in root
CLAUDE.md. It is the first thing a fresh session checks: what shipped, the date, what's now unblocked. Keep the accumulated model lessons — they're load-bearing. - Sweep
docs/problem.md— close what got resolved, add anything new that needs the owner. - Refresh the root
README.mdif the stage changed what the harness does or its numbers — its Status section hard-codes test counts and gate results and will otherwise silently lie to GitHub readers. - Tell it as a story (doc conventions): what was the problem → what did we do → why → what happened. Usually this lives in the design note's closing section and the commit message — not a separate report file.
- Commit as
docs: Stage N CLOSED — <one-line evidence>(match the precedent ingit log).
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.

