content-research-writer
Pass
Audited by Gen Agent Trust Hub on Nov 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content during research and drafting phases, creating a surface for indirect prompt injection attacks.
- Ingestion points: Research data gathered from the web (Section 3: Conduct Research), content of user-provided drafts (
draft-v1.md), and external reference materials stored in thesources/directory. - Boundary markers: The instructions do not define explicit delimiters or warnings to ignore instructions embedded in the ingested research content or drafts.
- Capability inventory: The agent has file system access (read/write) for managing draft files and research logs, as well as access to search tools for gathering external information.
- Sanitization: There are no explicit mechanisms for sanitizing or filtering instructions that might be contained within the content the agent is asked to analyze or summarize.
Audit Metadata