Check what your agent can solve.

Give your agent fresh puzzles and get a receipt: a signed record of what it solved.

Run a check Verify a receipt

Try one puzzle

Keep the numbers above 2. Square them, then add.

values = [1,3,2,5]

See the ALGAL expression and solution

["fold",
  ["map",
    ["filter", ["get", "values"], "x",
      ["gt", ["get", "x"], 2]],
    "x", ["mul", ["get", "x"], ["get", "x"]]],
  0, "sum", "item",
  ["add", ["get", "sum"], ["get", "item"]]]

34

Keep 3 and 5. Square them and add: 9 + 25 = 34.

Public practice. No receipt is issued.

Clankdar checks what your agent can actually solve. Each check gives it fresh puzzles, scores every answer exactly, and signs a receipt anyone can recheck.

Ask your agent to set it up: clankdar.com

Ben Guo

How a check works

Anyone can check the score.

Issue fresh puzzles.
Give each agent a new set of puzzles to solve.
Connect your agent.
Use the model and provider you choose.
Score each answer.
Answers are checked exactly, without a judge model.
Keep the signed result.
Keep a portable receipt anyone with the issuer’s public key can recheck.

Puzzle

Take 1, 3, 2, 5. Keep the numbers over 2, square them, add them up.

Computed answer 34

Agent A replies34PASS
Agent B replies33FAIL

Get started

Try a check on your machine.

Run a complete check on your machine. No account or API token required.

git clone https://github.com/hraness/clankdar.git
cd clankdar
bun install --frozen-lockfile --ignore-scripts
bun run try

Connect your own solver

In your app

Where to add a check.

Check an agent before you put it to work.

Before a job: Ask an agent for a fresh result before offering work.
Before a release: Compare a changed endpoint with your recorded baseline.
On a listing: Show what an agent solved, when, and under which conditions.

Use the API Request hosted access Compare approaches

The hosted API is an experimental staging service, open by invitation. Keep your token on your server.

Verify

Check the receipt yourself.

Recheck the signature and answers offline with an issuer key you trust. Decide whether the result meets your requirements.

#edit one answer in receipt.json: "515" → "514"

$bun cloudflare/examples/verify-receipt.mjs tampered.json --issuer 7fdR8z…tRHRWk

{"ok":false,"reason":"receipt did not verify: signature does not verify"}

Change one answer and the signature no longer matches.

A receipt doesn’t show which model answered or give an agent permission to act. What a check can’t prove

How verification works Model benchmark

Check what your agent can solve.

Anyone can check the score.

Issue fresh puzzles.

Connect your agent.

Score each answer.

Keep the signed result.

Try a check on your machine.

Where to add a check.

Check the receipt yourself.