Independently verify that a finished change does what was asked: rerun tests, build, lint, and exercise the behavior end to end before reporting done.
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# Verify before reporting done
1. Restate what "done" means for the request as 2-5 checkable statements.
2. Run the project's real checks, not a subset: the test command, type checker, linter and build
that CI runs (read `.github/workflows/*` to find them). Record exact commands and exit codes.
3. Exercise the changed behavior directly (see the `run` skill): one normal case, one edge case,
one failure case.
4. Review the full diff (`git diff` against the base) for leftovers: debug prints, commented-out
code, unrelated edits, secrets, TODOs you introduced.
5. If anything fails, fix it and repeat from step 2. Do not report success on static reading alone.
6. Report: each statement from step 1 with evidence (command + result), and anything you could not
verify and why.
The line at the top of this page pins the label git:20261029.2b2a16a, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
GET https://markdownregistry.com/api/v1/artifacts/art_bki33ifguemmxcsg
GET https://markdownregistry.com/api/v1/resolve?ref=fahthiz/oasis/verify
GET https://markdownregistry.com/api/v1/blob/73ae7a33de31e68fa19f1fdea34cd555ade4c08b67ad675cafb47ddbb9c8a5ff
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
fahthiz/oasis · evals/snapshots/oasis-main/skills/environment-setup/SKILL.md · Detect the environment and install or configure the smallest set of tools required by a task.
fahthiz/oasis · evals/snapshots/oasis-main/skills/ship-feature/SKILL.md · Implement a complete feature from requirement through tests and honest delivery notes.
fahthiz/oasis · plugins/oasis-research/skills/deep-research/SKILL.md · Multi-source web and code research with source verification, conflicting-evidence checks, and cited reporting. Use for…
fahthiz/oasis · skills/api-development/SKILL.md · Design and implement validated, authenticated APIs with clear error contracts and safe lifecycle handling.
fahthiz/oasis · skills/code-review/SKILL.md · Review current changes for correctness, security, performance, accessibility, regressions, and test gaps.
zenbu-labs/terminal-browser · .claude/skills/verify/SKILL.md · Drive an engine app headlessly in a pty, record a video of the whole verification, and open a summary page (video +…
zereight/gitlab-mcp · .github/skills/verify/SKILL.md · Verify that a change really works before claiming completion. Activate when user says: verify this, check if it works…
sethgammon/citadel · skills/verify/SKILL.md · Self-test the Citadel hook pipeline from within a live session. Exercises real tool calls (Write, Edit, Bash, Read) and…
