security-review-owasp-injection-prevention
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a persona and methodology for the agent to conduct security audits based on OWASP standards. It contains no executable code, network requests, or attempts to bypass agent safety guidelines.
- [INDIRECT_PROMPT_INJECTION]: The skill specifies a process for analyzing untrusted inputs and codebases. As an auditing tool, it inherently processes external content which could contain malicious payloads.
- Ingestion points: Data ingested includes code, configuration, tests, and documentation (SKILL.md).
- Boundary markers: None defined; data is processed directly as part of the review.
- Capability inventory: No tool invocations, file writes, or network operations are present in the skill instructions.
- Sanitization: No input sanitization is performed.
- [METADATA_POISONING]: The frontmatter metadata correctly identifies the skill's intent and parameters without deceptive content.
Audit Metadata