agentleFS
effectustasieffectustasi/agent-receipts/skills/prove-it/SKILL.md

prove-it

Skill1 starsChanged 41 days agoMITView on GitHub

MANDATORY before your final message on any coding task (bug fix, feature, refactor). Load it before you write done, fixed, works, or tests pass. Defines what counts as proof: fresh command output from this session that would fail if the change were wrong. Edited tests and happy-path runs do not count; say "not verified" when you have no proof.

27 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use root-cause, challenge and 4 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Ingramradical235Skill

    root-cause

    Structured root cause analysis using WHERE (MECE decomposition) then WHY (causal structure diagrams) pipeline. Includes coin-flip detection, logical-leap detection, and vicious cycle intervention. Use when buffer enters YELLOW/RED, when metrics stall, or when Actions execute but don't produce results.

    Also in 3 projects that use this

  • michalattSkill

    challenge

    Pressure-test plans, decisions, and directions by picking and answering the most relevant challenge questions for the current conversation

    Also in 3 projects that use this

  • renatorozasSkill

    grill

    Adversarial code review. Don't let me ship until the changes pass your scrutiny.

    Also in 4 projects that use this

  • AWSSkill

    retro

    Conduct a retrospective on a coding session.

    Also in 3 projects that use this

  • MicrosoftSkill

    handoff

    Compact the current conversation into a handoff document for another agent to pick up.

    Also in 3 projects that use this

  • GoogleSkill

    tdd

    Test-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 25 edited versions we read keep these.

  • 7

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After your last edit, run the smallest check that would fail if your change were wrong.

    line 27
  2. Distinctive:

    Read the output. Exit code, failures, and warnings that touch your change.

    line 28
  3. Distinctive:

    If you cannot verify (no test environment, needs credentials, needs hardware, would take hours), say exactly that.

    line 30
  4. Distinctive:

    Do not change its expected value to make it pass, unless the user asked for exactly that behavior change.

    line 59
  5. Distinctive:

    Look for a fix that passes the old tests and the new case.

    line 60
  6. Distinctive:

    If the two really conflict, stop and say so: name the test, the case, and the choice the user has to make.

    line 61
  7. Distinctive:

    Never: write a test that passes trivially just to have a receipt.

    line 68

“Common” is measured: kept by at least 6 in 10 of the 25 edited versions of this file we could read. “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

Prove it

You may not say a task is done, fixed, or working unless you have a receipt from this session.

What counts as a receipt

  • Output of a command you ran after your last edit: the tests, the build, the type checker, the script itself.
  • For UI changes: a screenshot or DOM read of the running app showing the change.
  • For a bug fix: the original reproduction now behaving correctly.

What does not count:

  • "The code looks right."
  • Tests you did not run, or ran before your last edit.
  • A passing test that never executes the code you changed.
  • A test you edited in this session so it passes with your change.
  • A run that never hit the failure: the happy path, or input that never triggered the bug.
  • Reasoning about what the output would be.

The rule

  1. After your last edit, run the smallest check that would fail if your change were wrong.
  2. Read the output. Exit code, failures, and warnings that touch your change. Not just the last line.
  3. Report with the receipt: the command and the lines that prove it.
  4. If you cannot verify (no test environment, needs credentials, needs hardware, would take hours), say exactly that. Never round "unverified" up to "done".

Report format

Verified:

Done. Receipt:
```
$ pytest tests/test_auth.py
5 passed in 0.41s
```

Not verified:

Changed `parse_date` to accept ISO weeks. **Not verified**: no test runner is set up here.
To check: `python -c "from app.dates import parse_date; print(parse_date('2026-W40'))"`

Phrases you may not use without a receipt

"should work", "this fixes it", "all tests pass", "verified", "confirmed", "Done ✅", "works now".

When an existing test fails after your change

Your change is the suspect, not the test. An existing test encodes behavior someone wanted.

  • Do not change its expected value to make it pass, unless the user asked for exactly that behavior change.
  • Look for a fix that passes the old tests and the new case.
  • If the two really conflict, stop and say so: name the test, the case, and the choice the user has to make.

Edge cases

  • Check failed? Say so first and show the failure. Do not bury it under a summary of what you changed.
  • Failures unrelated to your change? Name them. Only call them pre-existing if you showed they fail without your change too (for example with git stash); otherwise say you did not check.
  • Partial verification? Say which part is proven and which is not.
  • Never write a test that passes trivially just to have a receipt. The check has to exercise the change.

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.