Claron
    Documentation
    VERIFICATIONReviewed

    Human Verification

    Verification records whether a person checked a reported issue on a specific fixed version. Keep it separate from applying a patch, changing a status, or merging a pull request. Overview: review rounds and verification.

    Before you start

    A proposed fix on a known build or deployment, plus an eligible verifier.

    What you get

    A human approval or rejection, recorded against the selected version.

    Request verification

    1. Deploy or build the proposed fix and confirm the intended environment/version.

    2. Open the annotation's verification controls with permission to request verification.

    3. Choose an eligible verifier and the target build or deployment, then submit the request

      Verify that the request is pending for the intended person and target.

    4. The verifier can find the request in My Work and open the report's context.

    Approve or reject the result

    1. Open the assigned request and check the target build/deployment, expected result, and reproduction steps.

    2. Repeat the steps on that target

      Use the recorded evidence and Claron It where useful, but check the behavior itself.

    3. Approve when the intended result passes

      If it fails or you cannot verify the right target, reject and enter a concrete reason, including the failing step.

    4. Confirm the recorded decision and version

      The assigned verification task clears when its decision is recorded; merely reading the report does not complete it.

    If verification cannot proceed

    Check that you are the assigned eligible verifier, can access the target version, and have permission for the action. A pending status-approval request and a verification request are different workflows.

    For production extension issues, perform a human check on the chosen deployment. Automated retakes, automated after screenshots, and pixel comparisons are unavailable for live production annotations. Owners, developers, and testers can save native live after captures using manual after capture. A missing capture or a saved image alone must not be treated as a passing test.