Sensitive data in application logs
An AI assistant added 'helpful' logging to your TypeScript login function so you can debug auth issues. It compiles and the logs are useful. What should you catch on review, and why is sensitive data in logs a real problem even though logs feel internal?
Implement
redact_sensitive_log(line: str) → strExamples
in
["user=alice@example.com password=hunter2"]out"user=alice@example.com password=[REDACTED]"in
["Login attempt email=a@b.com sessionToken=abc123"]out"Login attempt email=a@b.com sessionToken=[REDACTED]"in
["status=ok latency_ms=42"]out"status=ok latency_ms=42"What a strong answer looks like
Treat the AI’s output as a draft to verify, not an answer to trust. Name the specific flaw and the input that triggers it, say how you’d catch it (tests, edge cases, reading critically), and how you’d re-prompt or decompose to get it right.
0:00 of about 11 min
Vibe & agentic: describe the solution in plain language (or narrate it) and the coach grades your approach.
Which questions mattered is sealed until you submit. Telling you now would just be handing over the edge cases.
Run or narrate your approach, then ask the coach.