content-research-writer
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill consists entirely of markdown-based instructions for the AI agent and does not include any scripts, binaries, or complex configuration files.
- [INDIRECT_PROMPT_INJECTION]: The skill maintains an attack surface for indirect prompt injection as it is designed to fetch and process data from external sources and user-provided drafts.
- Ingestion points: The "Conduct Research" task involves searching and extracting information from external web sources, and "Provide Section-by-Section Feedback" reads user drafts like article-draft.md.
- Boundary markers: There are no explicit delimiters or instructions to ignore potential commands embedded in the research material or drafts.
- Capability inventory: The skill possesses the capability to perform network searches and read/write files within the designated project folder.
- Sanitization: No specific sanitization or filtering logic is provided for the content retrieved during research before it is presented to the user or added to the draft.
- [SAFE]: The skill's operations are transparent and aligned with its purpose. The suggested file system usage is restricted to a dedicated folder and does not attempt to access sensitive system files or environment variables.
Audit Metadata