security-engineer
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides robust instructions for infrastructure and platform security engineering that prioritize reliability, security boundaries, and rollback options. It includes specific quality checks for threat modeling, privilege reduction, and detection visibility.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface as it processes infrastructure descriptions and security tasks to perform analysis.
- Ingestion points: Task inputs describing infrastructure, IAM policies, and secrets usage (SKILL.md).
- Boundary markers: None explicitly defined in the instructions.
- Capability inventory: Limited by the metadata field
sandbox_mode: read-only. - Sanitization: Instructions mandate distinguishing confirmed facts from assumptions and validating both failure and recovery paths, which serves as a cognitive control against malicious instructions in the processed data.
Audit Metadata