done
Skill11 starsChanged 14 days agoMITView on GitHub
Finish the current change by validating, committing, opening and merging a PR or MR, reconciling related issues, and updating the local default-branch worktree. Manual invocation only: use when the user says /done or $done.
37 other projects use a version of this.
Adopted by
Most recently taken up by
- GauravAlbal/agent-skillstheir own version, first seen Nov 12, 2026
- spongetimblin/claude-skill-librarytheir own version, first seen Nov 8, 2026
- vtmocanu/skillstheir own version, first seen Nov 7, 2026
- Melvynx/ai-maxxing-skillstheir own version, first seen Nov 1, 2026
- duthaho/skillhubtheir own version, first seen Oct 31, 2026
- bholmesdev/skillstheir own version, first seen Oct 18, 2026
Used with
Projects with it also use simplify, work and 4 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
Hugging FaceSkillsimplify
Refactor code after tests pass. The "Refactor" phase of Red-Green-Refactor.
Also in 5 projects that use this
vbrevikSkillwork
Use when the user says 'plan', 'todo', 'copy plan', 'append plan', 'resume plan', 'priorities', or 'what's next'.
Also in 3 projects that use this
OpenAISkilltranscribe
Transcribe audio files to text with optional diarization and known-speaker hints. Use when a user asks to transcribe speech from audio/video, extract text from recordings, or label speakers in interviews or meetings.
Also in 3 projects that use this
MicrosoftSkillhandoff
Compact the current conversation into a handoff document for another agent to pick up.
Also in 9 projects that use this
genomewalkerSkilllearn
Learn from web sources, generate SKILL.md, and store key patterns in soul memory. Trigger keywords: learn, skill-gen, web-learn, documentation, tutorial.
Also in 3 projects that use this
GoogleSkillgrill-me
A relentless interview to sharpen a plan or design.
Also in 4 projects that use this
Other versions37 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 37
- bholmesdev/hubble.md★ 1.5ksame name
- bholmesdev/skills★ 179same name
- Melvynx/ai-maxxing-skills★ 46same name
- endlessblink/master-plan★ 26same name
- fredoliveira/agents★ 24same name
- duthaho/skillhub★ 9same name
- Trejon-888/ix-claude-code-starter-kit★ 9same name
- HaipingShi/coderail★ 8same name
- MusserLab/lab-claude-skills★ 6same name
- ohm41321/luciazero★ 6same name
27 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository19 other files it gives its agents.
- AGENTS.mdAGENTS.md
- obsidian-skills/AGENTS.mdAGENTS.md
- add-just-doctorSkill
- agent-autoresearchSkill
- agent-harness-repairSkill
- change-evidenceSkill
- ci-speedSkill
- verifySkill
- deck-style-gallerySkill
- longform-fictionSkill
- loop-kitSkill
- loop-nightly-reviewSkill
- loop-triage-githubSkill
- loop-watch-socialSkill
- orch-claude-pi-codexSkill
- skills-doctorSkill
- tdd-pstackSkill
- teach-pstackSkill
- windows-leftover-cleanupSkill
The file
We split this skill into the rules it gives an agent
1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.
0
Common
Most of the 33 edited versions we read keep these.
1
Distinctive
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 11
Treat user arguments as optional scope, issue context, or merge-method preferences.
“Common” is measured: kept by at least 6 in 10 of the 33 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Done
Take the current task to an observable merged state. Manual invocation is explicit authorization to commit, push, create or edit a PR/MR, merge it, and mutate directly related issues. It does not authorize force-pushes, protection bypasses, unrelated edits, or other external writes.
Treat user arguments as optional scope, issue context, or merge-method preferences. Otherwise infer scope from the current conversation and repository evidence.
Process
Establish the closure contract. Resolve the git root, forge from the remote URL, authenticated user, default branch, current branch and upstream, worktree layout, dirty state, and any open PR/MR for the branch. Store the resolved forge default branch as
DEFAULT_BRANCHand use that value consistently below; never assume a fixed branch name. Define the intended change set from the current task; preserve unrelated work unstaged. If intended and unrelated edits cannot be separated safely, ask one precise question before any external write.Completion criterion: the forge, resolved
DEFAULT_BRANCH, intended paths, unrelated paths, existing review state, and local default-branch worktree are accounted for.Validate the intended change. Read the full diff and repository instructions. Run the smallest relevant checks after the final edit, including tests, lint, build, or generated-file checks required by the changed surface. Fix failures caused by the intended change; stop with the exact blocker when a green result requires broader scope or unavailable infrastructure.
Completion criterion: every intended change is represented in the diff, required artifacts are present, and no relevant check is red.
Commit and push. When on
DEFAULT_BRANCHor detached HEAD, create a short-lived branch such asdone/<short-slug>(or the prefix the repository documents) without disturbing the working tree. Follow recent commit-message style and split independent concerns into atomic commits. Stage only the intended paths, inspect the staged diff, commit, and push with upstream tracking. Resolve ordinary remote divergence without rewriting published history.Completion criterion: every intended change is in a remote-reachable commit and unrelated local changes remain untouched.
Reconcile issues when useful. Using the matching forge CLI or API, inspect open issues assigned to the authenticated user or unassigned. Judge relevance by shared behavior, root cause, and scope rather than keyword overlap.
- For a strong match fully resolved by this change, add the forge's closing reference to the PR/MR.
- For a strong match that remains open, add a non-closing relationship.
- With no match, create an issue only when the problem or follow-up is independently worth tracking. Skip issue creation for small fixes, routine cleanup, or changes with no meaningful follow-up.
Completion criterion: record one evidence-backed outcome: closing link, related link, newly created issue, or deliberate skip.
Create or update the PR/MR. Reuse the open PR/MR for the branch when one exists; otherwise create a ready-for-review PR/MR targeting
DEFAULT_BRANCH. Derive the title and body from the actual commits and diff. Include a concise summary, verification evidence, the chosen issue relationship, and any known residual risks.Completion criterion: one non-draft PR/MR represents the remote branch and its description matches the delivered change.
Reach the merge gate. Inspect mergeability, required checks, approvals, and unresolved review threads. Watch pending checks. Fix failures attributable to the change, push the correction, and re-run the same gate. If a remote CI lane fails and this manually invoked skill has explicit push/PR/merge authorization, first read the actual failure and record its run URL, failed lane, confirmed root cause, local equivalent validation result, and unresolved code/test/privacy/security risks in the PR/MR and final report. If the repository documents a standing CI-failure policy, that evidence permits proceeding without a second confirmation; if it does not, stop and report the failed lane before merging. Neither case permits hiding failures, bypassing protections, force-pushing, or ignoring non-CI merge blockers. 🔴 CHECKPOINT: before merging with any failed required lane, state the run URL, failed lane, root cause, local validation, and residual risk in the PR/MR and the final report. After merge, disclose the failed lanes, why the standing policy applied, and local validation status to the user. Respect repository protections and use the repository's established merge method.
Completion criterion: the PR/MR is mergeable, approvals are satisfied, and checks either pass or each failed CI lane has the required standing-policy evidence and risk disclosure; otherwise report the exact external blocker without claiming completion.
Merge and reconcile issue state. Merge with
ghorglab, then query the forge to verify the merged state and merge commit. Confirm that closing-linked issues are closed; close a still-open issue only when the merged change fully resolves it.Completion criterion: the forge reports the PR/MR merged and every touched issue has the intended final state.
Refresh the local default-branch worktree. Use
git worktree list --porcelainto find the worktree already onrefs/heads/$DEFAULT_BRANCH. If none exists and the current worktree is clean after the merge, switch the current worktree to$DEFAULT_BRANCH. Rungit pull --ff-onlyin the selected worktree. Preserve any unrelated dirty state and report it if it blocks the fast-forward.Completion criterion: the local
DEFAULT_BRANCHworktree is identified, its HEAD matches the latest remoteDEFAULT_BRANCH, and the merged change is reachable from it.
Final Report
Report the commit hash(es), PR/MR URL and merge result, issue action or skip reason, verification commands, and default-branch worktree path and HEAD. List any untouched local changes or external blocker explicitly.
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.

