building-cloudflare-mcp

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

• Third-party content exposure detected (high risk: 0.90). This skill instructs the agent to add and call remote MCP servers via arbitrary HTTP URLs (e.g., the "claude mcp add --url https://my-mcp-server.username.workers.dev" command and the MCP "tools/list"/"tools/call" endpoints), causing the agent to ingest and act on content returned by user-controlled/public worker endpoints, which is untrusted third‑party content that could enable indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs adding a remote MCP server at "https://my-mcp-server.username.workers.dev", which Claude will call at runtime to perform tools/call operations (i.e., execute remote code and control agent tool behavior), so this is a runtime external dependency that can directly control the agent.