exa-search
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions recommend configuring an MCP server by running
npx -y exa-mcp-server. This downloads and executes the official integration package from Exa, a well-known provider of AI search services.\n- [PROMPT_INJECTION]: The skill facilitates the retrieval of untrusted data from the public web via neural search, which creates a surface for indirect prompt injection.\n - Ingestion points: Data from the internet enters the agent context through the
web_search_exaandget_code_context_exatools documented inSKILL.md.\n - Boundary markers: No specific delimiters or instructions are provided to the agent to treat the search results as untrusted data.\n
- Capability inventory: The skill allows the agent to read arbitrary web content and code, which is then processed by the LLM.\n
- Sanitization: The skill does not include logic to sanitize or filter potentially malicious instructions contained within search results.
Audit Metadata