- Home
- Skills
- Testing & Debugging
- claim-check: no "done" until the checks pass
claim-check: no "done" until the checks pass
Makes your coding agent prove "done": runs the real typecheck, tests (and checks tests actually ran) and loads the app after redirects, then hands back the exact failures to fix.
Free
Works with the AI Agents you already use
claim-check: no "done" until the checks pass
Example session with this skill installed
I've fixed the login redirect and the auth tests. Run claim-check before you tell me it's done.
- Read your context and instructions
- Compiled the claim-check
Not done yet — two checks failed against the real project, so I'm not reporting this as finished.
What claim-check printed
Not done yet. These checks ran against the real project and failed:
1. [tests] The test command failed (exit 1). This is the real exit code, not a summary:
FAIL src/auth.test.ts
● login › rejects an expired token
2. [http_ok] The app returns HTTP 404 to a user opening "/". "/" redirects to /login,
which returned 404. The user cannot use it. If a redirect target is missing, create
that page — a redirect to a route that does not exist is a 404.
Fix these, then run this script again. Do not report the work as done until it passes.
What I did next
- Restored the expiry check in
src/auth.tsthat my change had removed, so expired tokens are rejected again. - Created the missing
app/login/page.tsxthat/redirects to. - Ran claim-check again:
claim-check: typecheck, tests, http_ok verified
Now it's done: typecheck, tests and the home page all pass on the real project.
Connects securely to your tools. The creator never sees your data.
About this skill
Coding agents end a turn with "tests pass" when no test ran, or "it works" when / redirects to a 404. claim-check makes the agent check before it claims: it runs the project's real commands and hands back failures it can act on.
What it checks
- typecheck:
tsc --noEmitmust exit 0. - tests: the test command must exit 0 and actually run at least one test. "0 tests passing" fails.
- http_ok: a browser opening
/must land on a 2xx after following redirects. A page that is still the starter template, or raw HTML with no stylesheet, also fails. - boot (optional): fails if no dev server is running.
How it works
- Before saying "done", "tests pass" or "it works", the agent runs
node scripts/verify.mjs .from the project root. - Exit 0: it reports which checks ran. Exit 1: it fixes every failure in the report and runs it again.
- A check that can't run (no tsconfig, no test script, no server) is skipped, never failed.
Always on in Claude Code
The hooks version — a Stop hook that won't let the turn end until the checks pass, plus a write verifier — is open source: github.com/mayank3354/claim-check
Built by NoCoder
The same checks gate every build in NoCoder (nocoder.codes), an AI app builder that builds full-stack web apps and tests them in a real browser before calling them done. Node 18+, no dependencies, MIT.
Requires Node 18+. Skips checks if tsconfig, test scripts, or server configurations are missing. Cannot verify logic beyond exit codes and HTTP status.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 2 days ago
- Free with an Agensi account
Requires