Skip to lesson
Exit
Debugging & Troubleshooting Method1 / 2

1 min lesson

Walk one all the way through

Make the call in this situation: "A user says the agent 'gives bad answers' on their large monorepo. Before blaming the model, what do you check first?" Explain what supports it.

Step 1 of 2

Walk one all the way through“agent gives bad answers”

  1. 1Clarify. "Bad" how - wrong code, ignoring files or refusing? Pin the exact behavior.
  2. 2Reproduce. Same repo, same prompt, repeated. Does it fail every time or intermittently?
  3. 3Isolate. Check the indexing indicator. Swap the model. Strip the repo down to the relevant files.
  4. 4Hypothesize. "Indexing never finished, so retrieval returns nothing and the agent answers blind."
  5. 5Test. Reindex, confirm completion, rerun the same prompt.
  6. 6Root-cause. Index stalled on a large or ignored directory - symptom was "dumb agent," cause was empty context.
  7. 7Resolve or escalate. If reindexing fixes it, write a KB note. If indexing stalls reproducibly, escalate with the repo and logs.
The pattern under all four

The user's words name a symptom; your job is to name the layer.

Interview move

Close every scenario with the same three beats: what you'd tell the user right now, the root cause in one sentence and whether you'd escalate and why. Most candidates trail off after the diagnosis. Landing all three reads as someone who's actually closed tickets.