creator-content-auditor
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local Python scripts (
rubric-score.pyandvalidate-audit-artifact.py) to process audit data and validate persistent artifacts. This is standard functionality for the skill's purpose as a scoring authority. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external content (influencer submissions, captions, and metadata). It identifies this as a potential attack surface and explicitly instructs the agent to treat such evidence as untrusted, which is a security best practice for data ingestion.
- [DATA_EXPOSURE]: The skill manages audit reports by writing them to a designated path (
memory/audits/influencer/). The instructions require the agent to ask for user permission before writing, ensuring human-in-the-loop control over file system modifications.
Audit Metadata