agent-native-architecture

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly exposes generic web/HTTP capabilities (e.g., WebTools.webSearch and webFetch in references/architecture-patterns.md and the generic call_api/fetch URL example in references/mcp-tool-design.md and from-primitives-to-domain-tools.md) and the required workflows/tests show the agent is expected to read and act on web/search results, meaning untrusted public third‑party content can be ingested and materially influence subsequent tool calls and decisions.