electronic-trusted-certificate

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

• Third-party content exposure detected (high risk: 1.00). The skill directly queries a user-supplied MCP HTTP endpoint (see SKILL.md default endpoint / scripts/http_mcp.mjs) and explicitly runs "list-tools" and parses JSON/SSE responses to choose tools and arguments, so untrusted third-party responses from that endpoint can influence which tool calls and actions the agent performs.