From atdd-kit
Use when explicitly invoked via /atdd-kit:express <issue> for trivial, documentation-grade changes (README edits, typo fixes, gitignore additions) with no functional breakage risk.
How this skill is triggered — by the user, by Claude, or both
Slash command
/atdd-kit:expressThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
For trivial, documentation-grade changes with no functional breakage risk.
For trivial, documentation-grade changes with no functional breakage risk. Skips PRD / User Stories / Plan / AT / structured review. Preserves Issue-driven governance and CI gate.
Error: Issue number required. Usage: /atdd-kit:express <issue-number>
gh issue view <N> to validate:
in-progress label → STOP: "Issue #N is already in-progress in another worktree."OK — express applies:
| Category | Examples |
|---|---|
| docs/README edits | Add a section, fix wording, update a link |
| typo fixes | Spelling errors in docs, comments, string literals (non-logic) |
| comment updates | Code comment clarity, inline doc corrections |
| gitignore additions | Add patterns, no logic change |
| version bump only | Bump a dependency version in a manifest with no behavior change |
NG — express does NOT apply:
| Category | Examples |
|---|---|
| New feature | Any new capability or endpoint |
| Behavior change | Altering logic, output format, or API surface |
| Dependency addition | New package or tool introduced |
| CI / hooks change | Modifying workflows, hooks, scripts |
| Security impact | Auth, permissions, secrets handling |
When in doubt: fall back to the full flow via /atdd-kit:defining-requirements <issue-number>.
This Issue appears to qualify for express mode (matched criterion: ). Proceed with express? [Y/n]
Do NOT start implementation until the user explicitly approves and states which OK criterion applies.
If the user declines, guide to /atdd-kit:defining-requirements <issue-number> instead.
Express does NOT create any intermediate artifacts under docs/issues/<NNN>/
(no PRD, no user stories, no plan, no acceptance-tests, no review report).
git checkout -b express/<N>-<slug><file>). Aborting express and switching to full flow."/atdd-kit:defining-requirements <issue-number> to restart properly.docs: <message> (#<N>)git push -u origin express/<N>-<slug>When the target repository is atdd-kit itself, the following are mandatory in the same PR — they may NOT be skipped:
.claude-plugin/plugin.json version bump (patch or minor per change type, per DEVELOPMENT.md)CHANGELOG.md [Unreleased] Added/Fixed/Changed entry describing the changeCreate the PR with:
gh pr create \
--title "<conventional-title> (#<N>)" \
--body "$(cat <<'EOF'
## Summary
<one-line description>
## Express Mode
**Applied criterion:** <which OK criterion was matched, with brief rationale>
**Scope confirmation:** Only documentation-grade files were modified. No logic, CI, or dependency changes.
Closes #<N>
EOF
)" \
--label "express-mode"
## Express Mode section is mandatory and must name the specific OK criterion that justified express (AC6).express-mode label is mandatory (AC5). If the label is not found, run /atdd-kit:setup-github first.Merge is performed by a human — this skill does NOT run gh pr merge automatically.
Once CI is green, notify the user: "CI passed. PR # is ready for your review and merge."
| Thought | Reality |
|---|---|
| "I'll start without user approval" | AC1: explicit user approval is mandatory before any implementation |
| "This one-liner touches logic — I'll finish and note it" | AC9: scope overflow → abort and report immediately |
| "CI is slow, I'll merge with --admin" | AC4: CI gate is non-negotiable |
| "I'll create the PR without the express-mode label" | AC5: label is mandatory for traceability |
| "I'll skip CHANGELOG / version bump for atdd-kit" | AC7: DEVELOPMENT.md rules apply even in express |
| "The Issue is already in-progress — I'll continue anyway" | AC3: in-progress collision → STOP |
npx claudepluginhub o3-ozono/atdd-kit --plugin atdd-kitGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.