audit-llm-security
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a template for performing security audits on other AI systems. While it mentions terms like 'ignore previous instructions' and 'jailbreak', these are documented as benign probes used for testing the resilience of a third-party application's security boundaries, not for attacking the agent itself.
- [SAFE]: The skill explicitly includes safety constraints for the agent, such as 'Stop at evidence; do not escalate into a weaponized jailbreak chain' and 'Never paste secret values, full system prompts, or live API keys into the report.'
- [SAFE]: The frontmatter and instructions define a structured, read-only process for inventorying AI stacks and identifying common misconfigurations (e.g., lack of human-in-the-loop for sensitive tools, missing output encoding, or unbounded token consumption).
Audit Metadata