ai-gateway

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

• Third-party content exposure detected (high risk: 0.90). This skill explicitly fetches and ingests remote third-party specifications/agent cards at runtime — see references/manifest-schema.md ("MCP Server (OpenAPI)" notes remote spec URLs are fetched and converted into MCP tools, and the "Agent" section warns that agent_card_url/hosted-a2a-agent sources are fetched and can influence agent behavior) which allows untrusted external content to materially change tool capabilities and agent actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches remote agent cards and OpenAPI specs at runtime (e.g., https://research-agent.example.com/.well-known/agent.json and https://api.weather.example.com/openapi.json), and those fetched documents are converted into agent tools/behavior that directly control prompts and capabilities, making them a high-confidence runtime risk.