deep-research
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the web, which could contain malicious instructions.
- Ingestion points: Search results from Perplexity and Kagi tools are used as core inputs in SKILL.md.
- Boundary markers: The instructions lack specific delimiters to isolate search results from the agent's logic.
- Capability inventory: The skill possesses Write, Task, and Read capabilities, increasing the impact of potential injection.
- Sanitization: No mechanisms for filtering or sanitizing external content are present.
Audit Metadata