Warning
This repository is deprecated and archived. Development has moved to ShiplightAI/agent-skills-v2. This repo is now read-only.
AI-powered test automation skills — ship with confidence by letting your coding agent verify, test, and iterate autonomously.
Single source of truth for every agent (Claude Code, Cursor, Codex, and 40+ more). Install skills with skills and the MCP server with add-mcp.
All skills depend on the Shiplight MCP server. Install it once for your agent (see Install below), or run:
npx add-mcp "npx -y @shiplightai/mcp@latest" -n shiplight --env PWDEBUG=console| Skill | Purpose |
|---|---|
/verify |
Visually confirm UI changes in the browser after a code change |
/create-yaml-tests |
Spec-driven E2E test creation: plan, scaffold, and write deterministic YAML tests by walking through the app |
/create-agent-tests |
Author and run coding-agent-driven Markdown test cases against a live environment (browser, API, DB, logs) with auditable PASS/FAIL reports |
/triage |
Reproduce failing E2E tests, diagnose root causes, fix YAML, and report app bugs |
/cloud |
Sync local tests with Shiplight cloud for scheduled runs, team collaboration (subscription required) |
| Skill | Purpose |
|---|---|
/test-coverage |
After a feature/PR, pick the testing strategy and drive the producers to write comprehensive tests; record what was tested and by what type |
/review is the single entry point. It triages your app and runs the right domain reviews, or jump straight to one with /review <domain> (e.g. /review security).
| Skill | Purpose |
|---|---|
/review |
Orchestrator + all review domains: security (OWASP, auth, injection, access control, supply chain), privacy (PII, consent, tracking, data flows, user rights), compliance (HIPAA, SOC 2, PCI-DSS, GDPR), design (visual, responsive, accessibility, typography, i18n), resilience (error handling, degradation, edge states, API contracts), performance (Core Web Vitals, bundles, runtime), SEO (meta, structured data, crawlability), GEO (AI citation readiness, llms.txt, entity clarity) |
npx -y skills add ShiplightAI/agent-skills -a claude-code -y && \
npx -y add-mcp "npx -y @shiplightai/mcp@latest" -n shiplight --env PWDEBUG=console -a claude-code -ynpx -y skills add ShiplightAI/agent-skills -a cursor -y && \
npx -y add-mcp "npx -y @shiplightai/mcp@latest" -n shiplight --env PWDEBUG=console -a cursor -yCursor disables newly-added MCP servers by default. Enable it: Cursor → Settings… → Cursor Settings → Tools & MCPs → Installed MCP Servers → shiplight (Disabled) — toggle the switch to enable.
npx -y skills add ShiplightAI/agent-skills -a codex -y && \
npx -y add-mcp "npx -y @shiplightai/mcp@latest" -n shiplight --env PWDEBUG=console -a codex -yPick from supported agents and swap the -a flag. Use --all to install to every detected agent, or -g for a user-level (global) install.
npx skills add ShiplightAI/agent-skills --allRestart your agent after installing.
npx skills add ShiplightAI/agent-skills --skill verify --skill triage -a claude-codeList all skills in this repo:
npx skills add ShiplightAI/agent-skills --listnpx skills updateThe /verify and /create-yaml-tests skills also include an opportunistic daily update check. When a coding agent uses one of those skills, the skill asks the agent to use .shiplight-agent-skills-last-update as the project-local attempt timestamp. On first use, if the file does not exist yet, the agent should create it and skip the update so a newly installed project does not pay for a second install. After that, the agent should run npx -y skills@latest update -y at most once every 24 hours. Treat this file as local cache and do not commit it.