Verify Before Done: Agent Skill (SKILL.md) for Claude Code, Codex and Cursor

Proves a task is finished before claiming it: restates the acceptance criteria, runs tests, typecheck, lint and build, exercises the changed path for real, checks for leftover to-do markers, debug logs and skipped tests, and reports evidence per criterion plus what wasn't verified.

Prompt

Installs to .agents/skills/verify-before-done-agent-skill-skill-md-for-claude-code-codex-an/

mkdir -p .agents/skills/verify-before-done-agent-skill-skill-md-for-claude-code-codex-an && curl -sL "https://prompts-gpt.com/prompts/verify-before-done-agent-skill-skill-md-for-claude-code-codex-and-cursor/raw" -o .agents/skills/verify-before-done-agent-skill-skill-md-for-claude-code-codex-an/SKILL.md
SKILL.md
Preview
---
name: verify-before-done-agent-skill-skill-md-for-claude-code-codex-an
description: "Proves a task is finished before claiming it: restates the acceptance criteria, runs tests, typecheck, lint and build, exercises the changed path for real, checks for leftover to-do markers, debug logs and skipped tests, and reports evidence per criterion plus what wasn't verified."
---

# Verify Before Done: Agent Skill (SKILL.md) for Claude Code, Codex and Cursor

Proves a task is finished before claiming it: restates the acceptance criteria, runs tests, typecheck, lint and build, exercises the changed path for real, checks for leftover to-do markers, debug logs and skipped tests, and reports evidence per criterion plus what wasn't verified.

## Prompt

```
---
name: verify-before-done
description: "Proves a task is finished before claiming it: restates the acceptance criteria, runs tests, typecheck, lint and build, exercises the changed path for real, checks for leftover to-do markers, debug logs and skipped tests, and reports evidence per criterion plus what wasn't verified. Use before saying a task is done, fixed or working, before committing or opening a PR, or when asked to 'double-check', 'verify' or 'make sure it works'."
---

# Verify before done

Before you say a task is done, fixed or working, prove it.

## Steps

1. Restate the acceptance criteria as a numbered list. If none were given, write the ones the request implies and say that you inferred them.
2. Run the project's checks with the commands the repo already defines (package scripts, Makefile, CI config), and paste the key lines of output (counts, failures, exit status), not the whole log:
   - tests: the affected ones, then the full suite if it's reasonable to run;
   - typecheck;
   - lint;
   - build.
3. Exercise the changed path the way a user would: run the CLI with real arguments, call the endpoint and show the response, open the page and check the changed state, run the migration on a local database. Record what you saw.
4. Map each criterion to the evidence that proves it. A criterion without evidence is not met.
5. Search the diff (`git diff` against the base branch) for leftovers: to-do or FIXME markers you added, debug logging, commented-out code, skipped or focused tests (.skip, .only, xit, fit), hard-coded test values, temporary files.
6. If a check fails and you think it was already failing, prove it: check out the base branch in a separate worktree (`git worktree add ../base-check <base-branch>`), run the same command there, then remove the worktree with `git worktree remove ../base-check`.
7. Report: each criterion as met or not met with its evidence; the checks run and their results; what you did not verify and why (no credentials, no browser, slow end-to-end suite); anything the user should check themselves.

## Rules

- Never write "should work", "should be fixed" or "I believe this works". Report the evidence or name the gap.
- If a check fails because of your change, the task is not done. Fix it or report the failure plainly.
- Don't weaken a test, skip a check or loosen a lint rule to get green unless the user agrees.
- Only run commands that are safe here. No deploys and no writes to shared data.

## Example output

```
1. Expired coupons rejected: met. "rejects expired" passes; POST /cart/coupon with code OLD10 returned 422 "error":"coupon_expired".
2. Valid coupons still apply: met. 13/13 existing coupon tests pass.
Checks: npm test 212 passed; tsc 0 errors; eslint 0 errors; build ok.
Leftovers: none found.
Not verified: receipt email wording (no mail sandbox in this environment).
```
```

## Usage Notes

Save as .claude/skills/verify-before-done/SKILL.md in one project (or ~/.claude/skills/verify-before-done/ for all of them); Codex reads .agents/skills/verify-before-done/SKILL.md. Then ask for the job in plain words and the agent loads it. Or install with: npx skills add prompts-gpt/skills --skill verify-before-done.

Example output

At a glanceVerification results for ledger export --sinceAll four acceptance criteria met, with evidence
Full suite tests passed
341
2 skipped, both pre-existing
Rows exported with --since 2024-03-01
37
Matches direct SQL count
Acceptance criteria met
4 of 4
Filter, no-flag, invalid date, boundary
Items not verified
3
Timezones, Windows paths, large exports

Example for: a Python CLI task, "add a --since date filter to ledger export so only transactions on or after that date are written to CSV."

Verification report: ledger export --since

Acceptance criteria
  1. ledger export --since YYYY-MM-DD writes only transactions dated on or after that date.
  2. Running without --since exports everything, unchanged from before.
  3. An invalid date exits non-zero with a clear message.
  4. The boundary date itself is included. (Inferred from "on or after".)
Checks run
CheckCommandResult
Affected testspytest tests/test_export.py18 passed, 0 failed
Full suitemake test341 passed, 2 skipped (both pre-existing, tagged @slow_network)
Typecheckmake typecheck (mypy)Success: no issues found in 47 source files
Lintmake lint (ruff)All checks passed
Buildpython -m buildBuilt ledger-2.4.0.tar.gz and wheel, exit 0
Exercised for real

Fixture samples/q1.db holds 120 transactions from 2024-01-02 to 2024-03-29.

$ ledger export --db samples/q1.db --since 2024-03-01 -o out.csv
Exported 37 transactions.
$ head -2 out.csv
date,amount,payee
2024-03-01,-42.10,Corner Grocer
$ ledger export --db samples/q1.db -o all.csv
Exported 120 transactions.
$ ledger export --db samples/q1.db --since 2024-13-01
error: --since must be a valid date (YYYY-MM-DD), got '2024-13-01'
$ echo $?
2

A direct SQL count of rows with date >= '2024-03-01' returned 37, which matches the export.

Criteria → evidence
  1. Filter applies: met. The export wrote 37 rows, matching the SQL count. test_since_filters_older passes.
  2. No-flag behavior unchanged: met. The export wrote 120 of 120 rows. All 11 existing export tests pass.
  3. Invalid date rejected: met. The command printed the error message and exited with status 2. test_since_invalid_date passes.
  4. Boundary inclusive: met. The first CSV row is 2024-03-01. test_since_includes_boundary passes.
Leftovers (git diff main...HEAD)
  • During development I added a print(rows[:3]) in export.py line 88. I have removed it, and rg "print\(" src/ledger/export.py now finds nothing.
  • No TODO or FIXME markers, .skip or xfail added, hard-coded dates outside tests, or temp files.
Not verified
  • Timezone edge cases. Dates are stored as naive strings. I did not test a database written with UTC timestamps, because none exists in the repo.
  • Windows paths for -o. No Windows environment was available here.
  • Large exports (100k+ rows). I did not benchmark these. The filter runs in SQL, so I don't expect a slowdown, but this is unmeasured.
For you to check
  • Confirm whether --since should also accept a full datetime. It currently rejects 2024-03-01T09:00.
Written by AI from this prompt. Your results will vary.

Save as .claude/skills/verify-before-done/SKILL.md in one project (or ~/.claude/skills/verify-before-done/ for all of them); Codex reads .agents/skills/verify-before-done/SKILL.md. Then ask for the job in plain words and the agent loads it. Or install with: npx skills add prompts-gpt/skills --skill verify-before-done.

Make it repeatable

Keep using this prompt without pasting it in every time.

Prompt details

By
Prompts-GPT editors
Published
How prompts are written and checked

Did this prompt work?

Keep exploring

More in AI Coding Agents