Agent Run Transcript Review: What Went Wrong and Why

Analyse a failed or odd agent run step by step.

Added October 15, 2026. Written with AI assistance and checked for clarity and placeholders; not reviewed by a lawyer, accountant, clinician or other professional. How these prompts are made

The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.

Review an agent run. The user's request: [USER_REQUEST]. The transcript of the agent's steps (thoughts or plans if visible, tool calls with arguments, results, final answer): [RUN_TRANSCRIPT]. What the user expected: [EXPECTED_OUTCOME].

Walk through the transcript step by step and identify where the agent first deviated from the intended path. Make a table with Step, What happened, Was it correct, Cause (unclear instruction, tool description, missing information, tool failure, reasoning error), and Fix. Check whether the final answer claims anything that the tool results do not support, and whether the agent reported failures honestly. Count the steps that were unnecessary. Suggest changes to the prompt, the tool descriptions, or the workflow, ranked by expected effect, and a test case to confirm the fix.

Do not blame the user. Redact sensitive data in the transcript before use.

Add a worked short transcript analysis: the user asked for 'last quarter's sales'; the agent used the calendar quarter instead of the fiscal quarter, called the right tool with the wrong dates, and reported success; the cause is a missing definition in the prompt, and the fix is to add the fiscal calendar and a confirmation of the date range.

Fill in before use

Optional: type your details and the prompt above updates. Nothing is saved or sent.

Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.