agent-surface-rollout
Pass
Audited by Gen Agent Trust Hub on Apr 28, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill facilitates the propagation of repository-wide instructions across multiple agent surfaces, creating a trust chain vulnerability where malicious content could be spread across the repository's AI configuration files.
- Ingestion points: Reads from 'skills/' directory and root instruction files (SKILL.md).
- Boundary markers: The workflow does not define specific delimiters or instructions to ignore embedded commands within the transferred content.
- Capability inventory: The skill has write access to multiple agent-facing surfaces including '.github/', '.windsurf/', and '.agents/skills/' (SKILL.md).
- Sanitization: There is no evidence of content validation or sanitization performed during the rollout process.
Audit Metadata