plan
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to research and synthesize information from workspace files and external web sources. This behavior exposes the agent to indirect prompt injection if those sources contain adversarial instructions.
- Ingestion points: Workspace file reading (SKILL.md Research step 2) and external research via URLs (SKILL.md Research step 4).
- Boundary markers: No explicit delimiters or safety instructions are mandated for processing retrieved data.
- Capability inventory: File system read/search, markdown file creation, discovery command execution (git, tests, builds), and user interaction tools (specified throughout SKILL.md).
- Sanitization: The skill lacks defined methods for sanitizing or filtering input from external sources.
Audit Metadata