research-agent
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to perform autonomous research by searching and synthesizing content from the web, which is an untrusted data source. This exposes the agent to indirect prompt injection, where a malicious website could host instructions designed to subvert the agent's logic or output.
- Ingestion points: Web search queries and external website content accessed via the
research_agenttool (SKILL.md). - Boundary markers: Absent. There are no instructions to the agent to distinguish between its own internal instructions and instructions possibly found within retrieved web data.
- Capability inventory: Performs web searches, synthesizes external data into reports, and generates data visualizations/charts.
- Sanitization: Absent. The skill does not describe any sanitization or validation of content fetched from the internet.
Audit Metadata