code-search-exa

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly instructs the agent to use the web_search_exa tool to fetch code snippets and documentation from public third-party sources such as GitHub and StackOverflow, and to read/interpret those results as part of its workflow (extracting and acting on snippets), which exposes the agent to untrusted, user-generated content that could carry indirect prompt-injection instructions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's MCP configuration calls the external endpoint https://mcp.exa.ai/mcp?tools=web_search_exa at runtime to fetch search results/snippets that are injected into the agent's context (used to form responses and code examples), so the fetched content can directly control prompts and deliver executable code.