d3fend-deceive
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is comprised solely of markdown files providing guidance on MITRE D3FEND tactics. It contains no code, scripts, or instructions that trigger automated execution.
- [EXTERNAL_DOWNLOADS]: The documentation mentions established cybersecurity tools and platforms such as Canarytokens.org, T-Pot, and Cowrie as resources for implementing deception environments. These references are standard for security practitioners and are presented neutrally without automated download or execution mechanisms.
- [INDIRECT_PROMPT_INJECTION]: While the skill advises on monitoring adversary interactions in honeypots (which constitutes an ingestion point for untrusted data), it provides high-level guidance only and does not include scripts or tools that would process this data, effectively neutralizing the risk within the scope of this skill.
Audit Metadata