From twake-guidelines
Use when writing commit messages, creating branches, or opening pull requests on Linagora/Twake/Cozy projects. Enforces Conventional Commits, one subject per commit (atomic commits), a structured PR workflow (branch, commit, push, gh pr create), lowercase PR titles under 70 chars, and Summary-only PR bodies with no invented motivation and no em dashes.
How this skill is triggered — by the user, by Claude, or both
Slash command
/twake-guidelines:twake-git-conventionsThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Apply these rules when committing code, writing commit messages, branching, or opening pull requests.
Apply these rules when committing code, writing commit messages, branching, or opening pull requests.
Use Conventional Commits:
type(scope): subject line in imperative mood
Body explaining WHY this change is needed, when the reason is
known and non-obvious from the diff. Wrap body lines at 72
characters.
BREAKING CHANGE: describe the break and migration path.
| Type | When to use |
|---|---|
feat | A new feature visible to users |
fix | A bug fix |
docs | Documentation only |
style | Formatting, whitespace, no logic change |
refactor | Restructuring without behavior change |
test | Adding or updating tests |
chore | Build, tooling, dependencies |
BREAKING CHANGE: footer with migration guidance.A commit must handle exactly one subject. Multiple concerns in a single commit make the change hard to review, hard to merge, and hard to revert.
If the commit subject needs the word "and" to describe the change, split the commit.
feat/short-descriptionfix/short-descriptionchore/short-descriptionmain / master.main must always be production-ready. Use feature flags for unreleased work.Pull requests must be as atomic as possible: one concern per PR, mirroring the atomic-commit rule at a larger granularity.
Every PR goes through the same mechanical steps:
git diff and git status to understand every change in scope.feat/..., fix/..., or chore/... matching the nature of the change.git push -u origin <branch>.gh pr create --title "..." --body "...".feat: ..., fix: ...) when it fits.Only one section: ## Summary, followed by concise bullets. Write like a human. Focus on why only when the reason is genuinely known.
Do not include:
—, –, --).The PR body describes only what the current diff changes. Nothing else.
CLAUDE.md, AGENTS.md, or prior conversations, not even for motivation.✅ feat(auth): add passwordless magic-link login
✅ fix(calendar): prevent duplicate events on timezone change
✅ chore(deps): upgrade react-query from 4.2 to 4.3
❌ Updated stuff
❌ fix: bug
❌ feat: add login and fix navbar and update deps
✅ feat: add passwordless magic-link login
✅ fix(calendar): prevent duplicate events on timezone change
✅ chore: upgrade react-query to 4.3
❌ Feature: Add Passwordless Magic-Link Login to the Auth Flow (too long, wrong case)
❌ WIP: stuff (vague)
## Summary
- Add `fetchMagicLink` helper that posts to `/auth/magic-link`.
- Wire the `/login` route to render `MagicLinkForm` when the feature flag is on.
- Add `magicLink` feature flag defaulting off.
## Summary
- Adds a new login feature as discussed with the CTO last week.
- Users have been asking for this for months, so this ships it.
- This rewrites the entire auth flow and also fixes several bugs.
## Test plan
- [ ] Login works
- [ ] Logout works
Problems with the bad example: invented stakeholder motivation, context from outside the diff, multiple concerns in one PR, a test plan section, a checklist.
npx claudepluginhub linagora/twake-guidelines --plugin twake-guidelinesProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.