⌁ REPAIRBENCH

NEBIUS SANDBOX × NVIDIA NEMOTRON

Submit a candidate to an
enforced validation protocol.

For integration engineers fixing API clients that return HTTP 200 but silently lose or corrupt data. Start with the supplied pagination case. NVIDIA Nemotron proposes a replacement; Nebius Sandbox runs the checks; the control plane binds the artifacts and enforces one correction maximum. Review the result and export its evidence.

Two proposals maximum. One contract. No automatic deployment.

Try the pagination repair Watch the uninterrupted live demo

01 / SET THE TASK

Checking availability

Start with the missing page.

  1. Run the supplied pagination case.
  2. Review the proposed function, named checks and stopping reason.
  3. Download the bound evidence for your review.

Checking whether new runs are available. Saved results are always a separate view.

Choose input

Follow every page. Stop safely.

Repair a client that returned HTTP 200 while silently dropping page 2 or looping on a cursor.

  • Empty pages are not the end
  • Duplicate items stay in order
  • Cursor cycles stop before a repeat
Two proposals maximum. One contract. No automatic deployment.

The control plane permits one correction after a completed, bound failure. A failed correction stops the workflow. Uncertain effects are not automatically repeated.

Free for reviewers; this uses project credits. Starting sends the function and checks to NVIDIA Nemotron and Nebius Sandbox.

02 / FOLLOW THE EVIDENCE

Ready when you are

Review the bound result.

No live run has started. The supplied pagination case is ready.

  1. Propose
  2. Sandbox test
  3. Correct once
  4. Stop

The trace will appear here.

Each stage will show what ran, what changed, and what the checks observed.

    Execution occurs remotely in Nebius Sandbox. Final comparison occurs outside that runtime. Candidate and observation runner currently share a process, so reported outcomes are not independently attested. The trace preserves completed, failed and uncertain stages separately.

    Security features

    Process isolation — remote execution, external comparison

    Generated code runs in Nebius Sandbox, outside the website's control plane and the development machine. The comparison plan remains outside the candidate runtime.

    The execution request supplies no inherited environment, requests disabled networking, and sets a 20-second runtime limit and 64 KiB limit per output stream.

    Boundary: candidate and observation runner currently share a Python process inside Sandbox. Independent comparison is implemented; independently trusted observation is the next boundary to establish.

    One correction. A finite repair protocol.

    One initial proposal. At most one correction after a completed, bound failure. The same contract governs both attempts. A second failure terminates the loop; uncertain effects are not automatically repeated.

    The application controls continuation, bounding iterative generation and token consumption. Attempt reservations constrain admission; they are not an account-wide billing guarantee.

    Cryptographic binding. Evaluate the identified artifact.

    SHA-256 digests identify the contract, proposal and uploaded files. Bundle snapshots, digest verification, upload byte counts and returned run identifiers reject payload changes and mismatched records at those checks.

    A changed contract creates a different evaluation. Hash binding establishes artifact consistency; it does not authenticate candidate-reported behavior or attest execution.

    Deterministic gates. Model confidence grants no authority.

    Application code checks proposal schema, function name and signature, result structure and run binding. Incomplete, malformed or unbound execution results receive no passing verdict.

    Passing declared checks permits review and export. Deployment remains a separate human decision.

    Help test the review workflow.

    Integration engineers: try the pagination case without coaching. Can you identify the original data loss, explain the final stopping reason and download the evidence? Send the point where you hesitated and whether the output would help your next code review.

    Send workflow feedback · Read the short test guide

    These demonstrations use fictional inputs. A structured external usability study has not yet been completed.