deepresearch-integrator
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes content from external deepresearch result files, which represents an indirect prompt injection surface.
- Ingestion points: Reads multiple source files from user-provided directories (e.g., SKILL.md references a research/ directory).
- Boundary markers: No explicit markers or 'ignore embedded instructions' warnings are defined for the file content processing.
- Capability inventory: The skill performs file system operations including reading, writing, and moving files (e.g., moving to a processed/ directory), and uses the TodoWrite tool.
- Sanitization: No validation or sanitization of ingested content is mentioned.
- Mitigation: The skill incorporates a strict human-in-the-loop workflow, requiring the agent to propose changes and wait for explicit user approval before making any modifications to the report or file structure.
Audit Metadata