write-a-skill
Write a new skill for a Cherry stage agent (Plan, Design, Preparation, Development, Deployment or Issues), or improve an existing one. Use when the user wants to turn a workflow they repeat into a skill, or asks to make, edit or review a SKILL.md.
- License 许可
- MIT
- 最近更新
- 2026-11-10
- Stars
- ★ 3
- Forks
- 0
01 前言 Front Matter
- name
- write-a-skill
- description
- Write a new skill for a Cherry stage agent (Plan, Design, Preparation, Development, Deployment or Issues), or improve an existing one. Use when the user wants to turn a workflow they repeat into a skill, or asks to make, edit or review a SKILL.md.
03 SKILL.md 正文
Write a skill
A skill is a folder with a SKILL.md: front matter with name and description, then the
instructions an agent follows. Cherry gives each stage agent its own skills.
Where it goes
- The user's own skills:
~/.cherry/skills/<stage>/<name>/SKILL.md, where<stage>isplan,design,preparation,development,deploymentorissues. That folder is outside the project, so Cherry will not let you write there: write the skill in the project at_project/tmp/skills/<stage>/<name>/and tell the user the one command that moves it:mkdir -p ~/.cherry/skills/<stage> && mv _project/tmp/skills/<stage>/<name> ~/.cherry/skills/<stage>/. - A skill for everyone who installs Cherry (only when working on Cherry itself):
packages/daemon/stages/<stage>/skills/<name>/, with a row inpackages/daemon/stages/SOURCES.md.
Pick the one stage whose agent needs it. A skill that helps several stages can be copied into each, but most belong to one.
Steps
- Ask what it is for, one question at a time: what the user does by hand today, when the agent should use the skill (the words or situations that should trigger it), and what a good result looks like. Read an example of the work if the user has one.
- Draft the description first. It is the only part the agent sees before choosing the skill, so it says what the skill does and when to use it, with the trigger words, in one or two sentences.
- Draft the body: the steps in order, each one checkable; the project's conventions it
must follow; what to do when something is missing; how to report the result. Keep it short.
Move long reference material (templates, checklists, examples) into separate files next to
SKILL.mdand link them, so they load only when needed. - Review it with the user. Read it back as the agent would: is every step something the agent can actually do with its tools? Does it ask before anything outward-facing (push, publish, sending to a service)?
- Finalize in the right folder and say how to try it: open that stage's tab; the skill is listed in the empty chat and in Settings, Stage agents.
Rules
- No
allowed-tools: Cherry removes it, and every command a skill runs still asks. - No secrets, tokens or personal data in a skill.
- Scripts are fine, but the skill must say what each one does, and it must not send data to other services without saying so.
- Never write hooks, settings or MCP servers: Cherry loads only the skill folder.
04 同仓库更多 Skills
verification-loop
Run a six-phase verification of a Claude Code session's work — build, type check, lint, tests with coverage, security grep, and diff review — then produce a PASS/FAIL verification report. Use when verifying work after completing a feature or refactor, before creating a PR, or when quality gates must pass.
e2e-testing
Playwright E2E testing patterns, Page Object Model, configuration, CI/CD integration, artifact management, and flaky test strategies. Use when writing Playwright tests, structuring page objects, or fixing flaky E2E runs in CI.
browser-qa
Run automated post-deploy UI verification with a browser automation MCP (claude-in-chrome, Playwright, or Puppeteer): console-error and Core Web Vitals smoke checks, form and auth-flow interaction tests, screenshot visual regression across three breakpoints, and axe-core accessibility audits ending in a SHIP / DO-NOT-SHIP verdict. Use when testing a deployed feature on staging or preview, before shipping frontend changes, reviewing a frontend PR, or checking responsive layout and accessibility.
ai-regression-testing
Regression testing strategies for AI-assisted development. Sandbox-mode API testing without database dependencies, automated bug-check workflows, and patterns to catch AI blind spots where the same model writes and reviews code. Use when adding regression coverage to AI-assisted code, or when the same model both wrote and reviewed a change.
story-lifecycle
Manage a file-based epic → story → sprint development loop without requiring an external project management tool. Creates and tracks epics, breaks them into user stories, assigns them to sprints, and drives implementation story by story. Use when starting a new feature initiative or when the team has no Jira/Linear/Taiga but needs structured delivery tracking.
intent-driven-development
Turn ambiguous or high-impact product and engineering changes into scoped, verifiable acceptance criteria before or alongside implementation. Use when a user asks to clarify a feature, define acceptance criteria, de-risk a security/data/migration/integration change, prepare implementation requirements for another agent, or make a complex request testable. Do not trigger for trivial edits, straightforward fixes, active debugging, code review, or implementation requests whose acceptance conditions are already clear unless the user explicitly invokes this skill.