gdpr-ccpa-privacy-auditor
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from
website_urlandpolicy_text(SKILL.md, skill.yaml) and processes it to generate compliance reports. While it uses these for technical analysis, there is an inherent risk that malicious instructions embedded in the scanned website or policy text could influence the agent's output. The skill lacks explicit boundary markers or sanitization instructions for this external content. - Ingestion points:
website_urlandpolicy_textinputs (skill.yaml). - Boundary markers: None specified for the untrusted content.
- Capability inventory: Performs cookie/script crawling and NLP policy extraction.
- Sanitization: No explicit sanitization or escaping of external content is mentioned.
- [PROMPT_INJECTION]: The skill contains a 'GLOBAL PROTOCOLS' section (SKILL.md) that attempts to override the agent's default operational behavior by mandating specific 'modes' (BUILD, INCIDENT, EXPERIMENT) and enforcing rigid execution patterns like the 'Iron Law of Execution'. While these are framed as framework protocols, they represent an attempt to constrain and redefine the agent's underlying logic.
Audit Metadata