ray-writer
Pass
Audited by Gen Agent Trust Hub on Aug 13, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run a local Python script (
scripts/article_check.py) to validate article metrics such as word count, paragraph length, and the presence of banned phrases. The script uses standard libraries and only performs read operations on Markdown files, with no network or system-altering capabilities.\n- [PROMPT_INJECTION]: The skill processes external data (web clippings and research notes), which represents a surface for indirect prompt injection.\n - Ingestion points: Reads from user-provided materials in
<vault>/30-腇料/and temporary draft locations.\n - Boundary markers: Utilizes YAML frontmatter for metadata separation; however, there are no specific instructions for escaping untrusted content during interpolation.\n
- Capability inventory: File creation within the vault and execution of the local
article_check.pyutility.\n - Sanitization: Output is validated by the mechanical check script for formatting and banned words, providing a layer of content oversight.
Audit Metadata