deep-research
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because its primary function involves fetching and synthesizing information from untrusted external web sources. An adversary could potentially influence the research report by placing instructions in content retrieved by the agent.
- Ingestion points: External research data retrieved via research queries in SKILL.md.
- Boundary markers: No specific delimiting or instruction-ignoring markers are documented in the skill instructions.
- Capability inventory: Subprocess execution of local scripts (scripts/research.py) to process queries and manage research state.
- Sanitization: No specific sanitization or filtering of the retrieved research content is described.
- [COMMAND_EXECUTION]: The skill instructions require the execution of a local script, scripts/research.py, to perform research tasks, check status, and list results. This is standard behavior for the intended use case.
- [EXTERNAL_DOWNLOADS]: The skill documentation recommends installing the 'httpx' library via pip. 'httpx' is a well-known, reputable Python HTTP client, and its use is consistent with the skill's documented purpose.
Audit Metadata