agentleFS
Sign inStart free
richardochristjiarichardochristjia/skills/make-fixture/SKILL.md

make-fixture

Skill2 starsChanged 7 days agoMITView on GitHub

Reads an app repo and writes its browser walkthrough fixture (`.harness/`) for a person to review and commit. Detects the stack, dev command and backend calls, writes a fake backend with canned answers, a Playwright driver that blocks every non-localhost request, start/stop scripts and a selftest, and runs the selftest until it passes. Never commits, never contacts a real backend, never copies tokens. Use when a repo has no fixture, when ship-plan or approve-plan says "no fixture", or for "make a fixture" or "set up the walkthrough".

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Never contact a real backend,: not to read a response shape either.

    line 26
  2. Distinctive:

    Never copy or print a token: Read no credential files; use fake values.

    line 27
  3. Very specific:

    Never edit app code: Only the fixture folder, plus a .gitignore line for its temp files if needed.

    line 28

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Make fixture

Tier L runs skip the browser walkthrough when the repo has no committed fixture. You write one. A person reviews and commits it.

Input

  • The repo, the current checkout.
  • Optional: the fixture folder, default .harness/ (or the FIXTURE: line of a tier.md).

Steps

  1. Read the repo. Find the stack, the dev command and port, and every backend call the app makes (base URLs, endpoints, auth flow). Read AGENTS.md / CLAUDE.md first.
  2. Write the fixture to CONTRACT.md: every required file, written for this app. No stack templates; the contract says what, you decide how.
  3. Fake each backend call with canned data in fake/, shaped like the real responses. Fake the login too, with made-up tokens.
  4. Run selftest and fix until it passes. Run it with the network blocked if you can.
  5. Stop, uncommitted. Report what was faked and what could not be: websockets, third-party OAuth, payments, anything else. Those become manual: criteria.

Never

  • Never commit in the target repo. Leave the files for a person to review.
  • Never contact a real backend, not to read a response shape either. Take shapes from the code, types and tests.
  • Never copy or print a token. Read no credential files; use fake values.
  • Never edit app code. Only the fixture folder, plus a .gitignore line for its temp files if needed.

Output

  • Files written, one line each.
  • Faked: the backend calls and what each answers.
  • Not faked: what could not be, and why.
  • Selftest: the command and its result.
  • To commit: "review .harness/, then commit it yourself".

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free