interview-article
Pass
Audited by Gen Agent Trust Hub on May 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run the
datecommand to synchronize content with the current date and a bundled Python script (scripts/check_body_length.py) to verify article length against specific requirements. - [SAFE]: The skill references absolute local directory paths (e.g.,
/Users/itwanger/Documents/GitHub/PaiAgent-one) to provide the agent with development context for its research tasks. These paths are consistent with the author's profile and intended use case. - [PROMPT_INJECTION]: The skill demonstrates an indirect prompt injection surface by ingesting untrusted data from
sucai.mdand web search results without utilizing strict boundary delimiters. This is identified as a low-risk architectural pattern common to content creation tools.
Audit Metadata