agentleFS
Sign inStart free
iwritec0deiwritec0de/app-dev/skills/verification/SKILL.md

verification

Skill3 starsChanged 7 months agoMITView on GitHub

This skill should be used before claiming any work is complete, before committing, before creating PRs, or before reporting task status. Enforces fresh verification evidence for all completion claims. Provides the verification gate methodology: run commands, read output, then report.

47 other projects use a version of this, 1 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use evidence-driven-writing, experiment-results-planning and 10 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.
  • Norman-burySkill

    evidence-driven-writing

    Use when writing or revising Introduction, Related Work, background, literature synthesis, or any section where references must drive claims

    Also in 3 projects that use this

  • Norman-burySkill

    experiment-results-planning

    Use when designing experiments, result tables, mock planning data, evaluation protocols, or results sections before real data are final

    Also in 3 projects that use this

  • +10 more from Norman-bury

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 39 edited versions we read keep these.

  • 5

    Distinctive

    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:

    NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE: If you have not run the verification command in the current session, you cannot …

    line 20
  2. Distinctive:

    Every completion claim must pass through this gate:

    line 24
  3. Distinctive:

    Before claiming work is complete in this project, run each applicable command and confirm clean output:

    line 34
  4. Very specific:

    Run these in order. A type error caught by tsc is cheaper to fix than one discovered during next build.

    line 44
  5. Distinctive:

    If a project uses a different test runner, linter, or build tool, substitute the appropriate commands.

    line 46
  6. Distinctive:

    If you catch yourself using these phrases, stop and run the actual command.

    line 59

“Common” is measured: kept by at least 6 in 10 of the 39 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

Verification Skill

Evidence before claims, always. NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION.

Iron Law

NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE. If you have not run the verification command in the current session, you cannot claim it passes. Prior knowledge, assumptions, and memory of previous runs do not count. Only fresh output from the current session constitutes evidence.

The Five-Step Gate

Every completion claim must pass through this gate:

  1. Identify — Determine which command proves your claim. "Tests pass" requires running the test suite. "No lint errors" requires running the linter. "It builds" requires running the build.
  2. Execute — Run the command freshly. Not from memory. Not from a previous session. Right now, in the current session.
  3. Read — Read the FULL output. Do not skim. Do not assume. If the output is long, read all of it. Errors can hide at the end.
  4. Verify — Confirm the output actually supports your claim. A test suite that prints results but exits with a failure code does not pass. A build that emits warnings about missing dependencies is not clean.
  5. Assert — Only after steps 1-4 are complete, state your claim and include the evidence. Quote the relevant output.

JS/TS Verification Checklist

Before claiming work is complete in this project, run each applicable command and confirm clean output:

CheckCommandWhat it proves
Type checkingtsc --noEmitNo type errors across the project
Testsnpx vitest run or npm testAll tests pass
Lintnpx eslint .No lint errors or warnings
Formattingnpx prettier --check .All files match formatting rules
Production buildnext buildBuild succeeds, catches SSR/hydration issues

Run these in order. A type error caught by tsc is cheaper to fix than one discovered during next build. A lint error is cheaper to fix before tests run.

If a project uses a different test runner, linter, or build tool, substitute the appropriate commands. The principle is the same: run the real tool, read the real output.

Red Flags

These phrases are signals that the verification gate has NOT been passed:

  • "should" — "Tests should pass" means you did not run them.
  • "probably" — "This probably works" means you did not verify it.
  • "seems to work" — "It seems to work" means you observed something informal, not a verification command.
  • "I believe" — "I believe the build is clean" means you are guessing.
  • Premature "Done!" — Claiming completion without showing command output.
  • "All good!" — Without evidence, this is an opinion, not a verification.

If you catch yourself using these phrases, stop and run the actual command.

When to Verify

  • Before every commit — Run the full checklist. Do not commit code that fails any check.
  • Before marking a task done — The task is not done until verification proves it.
  • Before creating a PR — The PR description should reference verification results.
  • Before reporting status — "Feature complete" means verified, not "I think I finished writing code."
  • Before delegating "completed" work — If you hand off work as complete, you are vouching for it. Verify first.

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