proactive-intelligence
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data from the web into the agent's context during autonomous search operations.\n
- Ingestion points: Snippets from web searches are retrieved via the
WebSearchTooland concatenated into aknowledge_blockin theanalyze_with_proactive_intelligencefunction (SKILL.md).\n - Boundary markers: The injected data is delimited by
<proactive_research>XML tags within the system prompt to help differentiate it from core instructions.\n - Capability inventory: The skill uses a web search tool to retrieve information and updates the system prompt, creating a vector where untrusted data influences the model.\n
- Sanitization: The implementation does not include any mechanisms for sanitizing or filtering the content of the search results before they are processed by the LLM.
Audit Metadata