commits
Skill518 starsChanged 57 days agoMITView on GitHub
Use when picking a branch name, staging files, or writing a commit message — covers branch conventions, the project's Conventional Commits format with the scope vocabulary observed in history, and the documentation-update gate that fires before every commit (`docs/*`, `README.md`, godoc, Helm values).
We split this skill into the rules it gives an agent
3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.
0
Common
Most of the 8 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 51
Before staging the commit, walk the diff and update the matching docs in the *same* commit (or at minimum the same PR — never a follow-up).
- Very specific: line 76
Don't include [skip ci] in human or agent commits.
- Very specific: line 77
Don't --amend after a hook failure — the commit didn't happen, so amend modifies the *previous* one.
“Common” is measured: kept by at least 6 in 10 of the 8 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
Commits
Branches
| Use | Form |
| Feature work | feature/<short-slug> |
| Bug fix tied to an issue | fix/<NN>-<short-slug> (NN = issue number) |
| Codex / agent branches | codex/<slug> |
| Release / deploy | bot-owned — never push manually |
Branch from main, not development (no development branch in this repo). Keep one branch focused on one change. If you find yourself fixing two unrelated things, that's two branches.
Commit messages — Conventional Commits
<type>(<scope>): <imperative subject, lowercase, no trailing period>
<optional body — explain *why*, not *what*. Wrap at ~72 cols.>Types observed in history: feat, fix, chore, docs, test, perf. The repo doesn't use refactor, style, or ci as types — ci is a scope.
Scopes (use one of these when it fits; invent sparingly):
| Layer | Scopes |
| Backend domain | api, auth, oidc, db, metadata, indexer, downloader, importer, recommender, scheduler, notifier, series, author, abs, calibre, telemetry, opds, httpsec |
| Cross-cutting | ci, release, deploy, lint, deps / deps-dev, migrate, sync, scanner, search |
| Other | test, docs, web |
Issue link in the subject: append (#NN) — e.g. fix(indexer): strip possessive author prefix from title before relevance matching (#446).
Reserved: chore(deploy): promote bindery to vX.Y.Z [skip ci] is bot-only. Never include [skip ci] in human or agent commits — it bypasses CI and CI is the gate.
Subject examples from real history:
feat(oidc): test discovery button surfaces IdP errors before login (#460)fix(recommender): don't gate author-new / series / genre-popular on ratings countperf(author): parallelize author work discovery and searchestest(downloader): cover download client edge-case matrix (#451)docs(series): document enhanced hardcover series workflow
Documentation-update gate
Before staging the commit, walk the diff and update the matching docs in the same commit (or at minimum the same PR — never a follow-up).
| If your diff touches… | Update… |
| Env vars, config, startup flags, upgrade path | docs/DEPLOYMENT.md |
| Auth modes / OIDC / proxy auth | docs/auth-multiuser.md, docs/auth-oidc.md, docs/auth-proxy.md, docs/troubleshooting-auth.md as applicable |
| ABS import flow | docs/abs_import.md (+ docs/ABS-Import-Wiki.md for user-facing changes) |
| Hardcover / series feature | docs/Hardcover-Series-Wiki.md |
| Upgrade-from-v1 path | docs/upgrade-v2.md |
| Multi-user behaviour | docs/multi-user.md |
| Feature surface advertised at top level | README.md |
| Helm chart values, env, image config | charts/bindery/values.yaml |
| New exported Go symbol with non-obvious behaviour | godoc comment on the symbol |
| Agent process / lifecycle / skill triggers | AGENTS.md (only when the agent contract genuinely changes — don't routinely touch) |
Rule of thumb: the entire docs/ directory and README.md are maintained at every commit — they should never lag behind the code. AGENTS.md is updated only when contributor / agent process actually shifts. CONTRIBUTING.md is human-maintained — don't auto-edit unless the user asks.
Pure refactors with no behaviour change require no doc update.
Out of scope (release-time only)
CHANGELOG.md sections are authored at tag/deploy time, not at commit time. Don't add an [Unreleased] entry; don't modify versioned sections. The tag-release skill covers CHANGELOG authoring when cutting ## [vX.Y.Z].
Don't
- Don't include
[skip ci]in human or agent commits. - Don't
--amendafter a hook failure — the commit didn't happen, so amend modifies the previous one. Re-stage and create a new commit.
Other versions10 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 10
- JetBrains/intellij-community★ 21ksame name
- redis/RedisInsight★ 8.9ksame name
- DetachHead/rebased★ 6ksame name
- JetBrains/MPS★ 1.7ksame name
- portabletext/editor★ 280same name
- Aurealibe/claude-config★ 68same name
- ecosyste-ms/claude-plugin★ 3same name
- ecosyste-ms/skills★ 2same name
- mamachanko/skills★ 1same name
- tranviet0710/agent-skills★ 0same name
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.
Used alongsideProjects with it also use writing-tests.
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.
JetBrainsSkillwriting-tests
Write IntelliJ JUnit 5 tests with fixtures, lifecycle, and EDT.
Also in 3 projects that use this
In this repository10 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- autofixSkill
- discord-reportsSkill
- fix-issueSkill
- prsSkill
- smoke-testingSkill
- tag-releaseSkill
- testingSkill
- triageSkill
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.

