matz-psychological-targeting
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied audience goals and context to generate structured decision briefs. This establishes a data ingestion surface where untrusted input could attempt to influence the agent's output. However, the skill provides extensive boundary markers (M1-M12 rules) and lacks dangerous capabilities like shell execution, file writing, or network sending, effectively mitigating this risk.
- [EXTERNAL_DOWNLOADS]: The skill references several external URLs in its derivation metadata and source citations. These links point to reputable academic platforms (arXiv, Europe PMC), university magazines (Columbia University, University of Toronto), and professional expert sites. These references are used for source attribution and evidence verification rather than runtime code execution or software dependency management.
Audit Metadata