hunt-sqli
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is a set of instructions for vulnerability research, containing standard industry techniques for auditing databases and APIs. The content is purely educational and procedural for a security professional.
- [INDIRECT_PROMPT_INJECTION]: The skill describes tasks that involve processing untrusted data from target systems. It proactively includes defensive instructions to distrust hints in target responses, mitigating the risk of the agent being manipulated by deceptive content during the hunting process. 1. Ingestion points: Target HTTP responses and headers captured during vulnerability testing. 2. Boundary markers: Explicit instructions to distrust embedded text and hints on target pages. 3. Capability inventory: Use of security tools such as curl, sqlmap, and grep. 4. Sanitization: Relies on automated tools and specific proof criteria rather than interpreting target content as instructions.
Audit Metadata