Why AI-Assisted Bug Hunting Needs Guidance
· 0:59 · Watch on YouTube ↗
Summary
A speaker says AI-assisted bug finding works better with iterative steering: tailored agent files, harnesses, guardrails and prompts that specify what to find and what to avoid.
Key points
- The speaker says using AI for bug finding effectively requires guidance rather than relying on out-of-the-box behavior.
- They describe revising agent files and adjusting the harness and guardrails when the agent goes off track.
- Prompts can specify what findings to report, what to look for and what to avoid.
- The speaker suggests, as a hypothesis, that some users reporting worse results may give Claude only a broad request to find a bug on hackerone[.]com.