exa-web-search-free

Pass

Audited by Gen Agent Trust Hub on Nov 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The instructions direct the configuration of the mcporter tool to a remote endpoint at https://mcp.exa.ai/mcp. This endpoint hosts the tool logic for the search functionality which is common for MCP-based skills.
  • [INDIRECT_PROMPT_INJECTION]: The skill retrieves snippets, titles, and URLs from external websites, creating a surface for embedded instructions in web content to influence the agent.
  • Ingestion points: The results output field processes data retrieved from the Exa search engine in response to the search queries.
  • Boundary markers: No specific delimiters or instructions are provided to the agent to treat search results as untrusted content or to ignore potential instructions within the metadata or snippets.
  • Capability inventory: The skill utilizes the mcporter binary to invoke remote MCP tools for code and company research.
  • Sanitization: There is no evidence of filtering, escaping, or validation for the data returned from the search service before it is presented to the agent context.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 13, 2026, 09:39 AM