The Agent Skills Directory

[SAFE]: The skill defines a legitimate workflow for creative review. It contains no evidence of prompt injection, data exfiltration, or malicious persistence mechanisms.
[INDIRECT_PROMPT_INJECTION]: The skill contains a vulnerability surface as it ingests untrusted user content. 1. Ingestion points: The skill reads the 'exact material being tested' as defined in SKILL.md. 2. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present. 3. Capability inventory: The skill specifies writing notes to '00_project/reviews/', but contains no network access or command execution capabilities. 4. Sanitization: No sanitization of ingested content is mentioned. The lack of dangerous capabilities makes this surface safe.

audience-sim