onboard
Use when the user explicitly invokes /onboard or asks to "give me a tour of this codebase" / "I'm new here, where do I start?" / "summarize this project for me." Performs a structured codebase tour, infers conventions from the existing files, and proposes a CLAUDE.md (or updates an existing one) so subsequent sessions have working context.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-09
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- onboard
- description
- Use when the user explicitly invokes /onboard or asks to "give me a tour of this codebase" / "I'm new here, where do I start?" / "summarize this project for me." Performs a structured codebase tour, infers conventions from the existing files, and proposes a CLAUDE.md (or updates an existing one) so subsequent sessions have working context.
- disable-model-invocation
- true
03 SKILL.md 正文
Onboard
You are giving a new contributor (human or future-Claude session) a structured tour of an unfamiliar codebase. The output is both a tour-narrative and an artifact (CLAUDE.md) future sessions can read.
Process
1. Survey the layout
# Top-level shape
ls -la
# Detect monorepo / multi-project
find . -name "package.json" -not -path "*/node_modules/*" | head -20
find . -name "pyproject.toml" -not -path "*/.venv/*" | head -10
find . -name "Cargo.toml" -not -path "*/target/*" | head
find . -name "go.mod" | head
# Existing docs
ls README* CLAUDE.md AGENTS.md CONVENTIONS.md adr/ 2>/dev/null
2. Identify the stack
Per project (or per sub-project in a monorepo):
- Language(s) and version (from
engines/python_requires/ etc.) - Framework (React, Express, FastAPI, Rails, etc.)
- Test runner
- Lint/format tools
- Build system
3. Find the "story shape"
What does this code DO at a one-paragraph level? Read:
- The top of the main README.
- The 2-3 most recent merged PR descriptions (
git log --oneline -20). - The largest source file (often a hint at where complexity lives).
4. Detect conventions
Without writing anything yet, infer:
- Naming conventions in use (camelCase / snake_case).
- Test file location convention (sibling vs
__tests__/mirror). - Error-handling style (exceptions / Result-types / nulls).
- Comment style (sparse / heavy JSDoc / docstrings).
- Anti-patterns the codebase already avoids or doesn't.
5. Surface what's missing
- Is there a
CLAUDE.md? If not, propose one. - Is there a
CONVENTIONS.md? If not, suggest scaffolding from the plugin's template. - Is there a
REMEDIATION.md? If not, propose surveying for known bugs / hot files. - Is the lint/test setup detectable? If not, suggest
/init-guardrails.
6. Write the tour
Show the user a structured tour:
# Codebase tour: <project name>
## What this is
[One paragraph from the survey.]
## Layout
[Tree-style; just the meaningful directories.]
## Stack
| Layer | Tech |
|---|---|
| Frontend | ... |
| Backend | ... |
| Database | ... |
| Tests | ... |
## How to run
[The 3-5 commands a new contributor needs.]
## Where the complexity lives
The 2-3 files / modules new contributors should read first. Any
known fragile zones.
## Conventions in use
[Short list inferred from grep + sample reading.]
## What's missing (suggested next steps)
- [ ] CLAUDE.md (would help future Claude sessions)
- [ ] CONVENTIONS.md (suggested scaffold via /sync-rules later)
- [ ] CI workflow
- [ ] REMEDIATION.md (if known bugs)
7. Offer to commit the artifact
If the user wants the tour persisted, save as CLAUDE.md (or
CLAUDE.md + docs/codebase-tour.md if a CLAUDE.md exists). Do NOT
overwrite an existing CLAUDE.md without asking.
Constraints
- Read-only survey. Do not edit, install, or build during onboard.
- One pass. If the user asks "tell me more about X," that's a follow-up question, not another full tour.
- Don't fabricate. If you can't tell what a directory is for from reading, say "(unclear from contents — ask the team)".
- Time-box: 5-10 file reads should be enough. Reading every file is the wrong shape.
04 同仓库更多 Skills
test-first-extract
Use when the user asks to EXTRACT or RELOCATE existing code into its own file, module, or component ("extract X", "split out Y", "pull Z into its own file", "decompose god-file"). NOT for net-new features (use tdd-feature-build) or behavior changes (use fix-bug). Enforces the 9-step extraction with isolated unit tests + integration regression tests as a two-tier safety net.
tdd-feature-build
Use when the user describes a NEW feature to add ("add X to Y", "implement Z", "build a feature that…", "make it possible to…"). Enforces TDD — failing test first, minimum code to pass, refactor. Refuses to ship code without a test that demonstrably catches its absence. Does NOT trigger on bug fixes (use bug-fix-discipline) or refactors (use test-first-extract). Will ask 1-3 clarifying questions before writing any code; safe to delegate from /feature command.
sync
Multi-machine git-backed sync of local TDD Pro state via the `tdd-pro-sync` branch. Use when operators run claude-tdd-pro on more than one machine and want their FAILURE-LOG, decisions, audit checkpoints, workflow state, operator-curated URL registries, operator/community trees, and paywalled attestations to follow them between hosts.
strict-component-tests
Use when writing or reviewing test files. Enforces strict assertions that actually catch regressions (getByRole, exact counts, jest-dom matchers) in contrast to permissive tests (queryAllByText length > 0) that pass against broken UIs. References QUALITY-BAR.md for the canonical strictness rules.
spec
暂无简介
spec-first
Use when the user describes a non-trivial feature or system that should be specified in writing before any code is written. Pattern from Geoffrey Huntley's "Ralph" loop, OpenSpec, intent-driven.dev, and Kent Beck's "augmented coding" framing — specs+tests are the durable artifact, code is the disposable one. Produces a Markdown spec document that the user reviews; only after approval does any code get written (typically via tdd-feature-build or /feature).