DEBUGGING / A WORKED STARTING POINT
Separate the observation from the hypothesis
Turn an error report into a small discriminating check without presenting a guess as a confirmed cause.
Template reviewed
WHEN TO USE IT
You have a reproducible error but have not yet established the right fix.
Bring these inputs
- {{evidence}}
- The exact error, relevant input, code and any known contract.
01 / THE PROMPT
A template you can inspect.
Help diagnose the failure below using only the supplied evidence.
Return three short sections: Observed, Possible explanation, and Next check. Keep the observation separate from your hypothesis. Propose the smallest check that would distinguish the leading explanation from another plausible one. State what result would support or weaken it. Do not implement a fix, invent a run or assume an unstated data contract. If the intended behaviour is missing, ask one precise question about it.
Evidence:
{{evidence}}
Copying does not run the prompt. Workbench opens an editable draft with the example inputs. It does not save or send it.
02 / THE EXAMPLE TARGET
What a useful result should contain.
This is a target for the worked example, not a recorded model response. Equivalent wording can be valid where the task allows it.
Observed: int() rejected the string '12.0'. Possible explanation: the feed now represents whole counts using a decimal string. Next check: inspect the feed contract and neighbouring raw values. Ask whether decimal-form strings are allowed. Do not silently round or discard fractional values as a proposed fix.
03 / JUDGE THE RESULT
Check the answer, not the confidence.
- The actual rejected value is preserved.
- The feed-format change is a hypothesis, not a confirmed event.
- The next check resolves the contract before recommending coercion.
The worked example is a target to inspect, not a saved response from a model. Use the checks to judge an actual result.
This template was revised during the September prompt review. The earlier text remains in Git history.
Read the prompt collection review ↗