Agent skills for working with testRigor, the AI-based, plain-English test automation tool. Each subfolder is a self-contained skill — a SKILL.md plus any references — written as plain Markdown so any LLM coding agent can use it: Claude Code, Codex, Cursor, openclaw, and others.
| Skill | Use it to… |
|---|---|
testrigor-dev-loop |
Start here. The end-to-end playbook: build tests + reusable rules as files → run against localhost via the tunnel → debug from the JUnit report → update the remote suite by re-running. |
testrigor-write-tests |
Author test cases in testRigor's plain-English language, the on-disk file formats for test cases / reusable rules / variables, and API steps. Full command catalog in reference.md; API steps in api-testing.md. |
testrigor-cli |
Run/trigger suites with the testrigor CLI — auth, the file-mutation (update-remote) model, localhost tunnel, debugging signals, labels, JUnit, mobile builds. |
Typical flow (all orchestrated by testrigor-dev-loop): build files (testrigor-write-tests) → run on localhost and debug (testrigor-cli) → update remote by re-running.
Each skill is a SKILL.md whose YAML description says when to use it, with heavy detail split into companion files (reference.md, api-testing.md) that are read only when needed — so the base instructions stay small. How an agent picks them up depends on the agent:
- Agents that support the Agent Skills format (e.g. Claude Code) auto-discover the skills and load the relevant one when your request matches its
description— asking to "write a testRigor checkout test" pulls intestrigor-write-tests; "run the suite against localhost" pulls intestrigor-cli. You don't have to name a skill, but you can nudge by mentioning the capability. - Other agents (Codex, Cursor, …) use the same Markdown by pointing the agent at this repo, or referencing the
SKILL.md/reference.mdfiles from the agent's own rules/context mechanism (Cursor rules, anAGENTS.md, included context, etc.).
git clone https://github.com/TestRigor/skills.gitClaude Code — copy the folders into its skills directory; they're auto-discovered:
# personal — available in every project
mkdir -p ~/.claude/skills
cp -r skills/testrigor-* ~/.claude/skills/
# …or project-scoped (only inside one repo):
# cp -r skills/testrigor-* <your-project>/.claude/skills/Codex / Cursor / other agents — keep the cloned repo in (or alongside) your project and reference the skill files from your agent's instructions: e.g. point a Cursor rule or an AGENTS.md at testrigor-dev-loop/SKILL.md as the starting point, or just open the repo so the agent can read the files on demand.
The testrigor-cli skill needs the CLI (Node.js ≥ 18) and a Personal Authentication Token (PAT). Get a PAT in the app: username (top-right) → API Tokens → Generate New Token.
npm install -g testrigor-cli
export TESTRIGOR_API_KEY="<YOUR_PERSONAL_AUTH_TOKEN>" # non-interactive (best for agents/CI)
testrigor authenticatealso stores a token, but it prompts interactively — fine for a human to run once during local setup, but an agent/CI loop should use theTESTRIGOR_API_KEYenvironment variable (or--auth-token) so nothing hangs.
testRigor also ships an MCP server — the agent-native alternative to the CLI for running suites and reading results as structured data (no JUnit parsing), with no interactive prompts. Add it to Claude Code with your PAT:
claude mcp add --transport http testrigor https://api2.testrigor.com/api/v1/mcp \
--header "personal-access-token: <YOUR_PAT>" -s userOther MCP-capable agents (Cursor, etc.) point their own MCP config at https://api2.testrigor.com/api/v1/mcp with the same personal-access-token header. The server can list/retrieve suites & cases, run test cases or a green regression, list runs, and read run failures. Guide: https://testrigor.com/how-to-utilise-testrigors-mcp-server/
Two complete, runnable sample projects ship inside the testrigor-dev-loop skill (so they travel with it on install), under testrigor-dev-loop/examples/ and documented in testrigor-dev-loop/examples.md — simple-search for the basic loop and shop-checkout for the localhost tunnel with labels, variables, and negative tests. The testrigor-write-tests skill additionally bundles examples.md, a worked authoring mini-project covering every file type. Run the projects to watch the build → run → debug → sync loop end to end.