gcp
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- [Prompt Injection] (SAFE): No instructions found to override system safety or bypass restrictions.
- [Data Exposure & Exfiltration] (SAFE): No patterns detected for accessing sensitive files or exfiltrating data to external domains.
- [Remote Code Execution] (SAFE): No remote script downloads or unverified dependency installations are present.
- [Command Execution] (SAFE): While the skill provides high-privilege commands, it mandates explicit safety warnings and backup reminders for destructive operations.
- [Indirect Prompt Injection] (SAFE): The skill processes user input to generate technical advice. Evidence Chain: 1. Ingestion point: User queries (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: CLI/YAML generation for user review (SKILL.md). 4. Sanitization: Absent. The risk is negligible as it functions as an advisor without autonomous execution capabilities.
Audit Metadata