exa-search

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.95). The skill’s required workflow uses Exa MCP tools like web_search_exa / web_search_advanced_exa / get_code_context_exa / crawling_exa, which fetch and ingest arbitrary public web/code content at runtime (outsider-authored free text) into the agent’s LLM context.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill requires running "npx -y exa-mcp-server", which fetches and executes remote code from the npm registry (e.g. https://registry.npmjs.org/exa-mcp-server) at runtime, so an external package controls/executes code the agent depends on.