mcp-use-code-review

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

• Third-party content exposure detected (high risk: 0.90). The skill's workflow explicitly connects an MCPClient to remote MCP servers via config "url" entries (see Server Config / references/server-config-patterns.md and deployment examples) and then discovers/reads tools, prompts, and resources with session.list_tools(), session.read_resource(), and agent.run(), meaning untrusted third‑party server-provided content is ingested and can directly influence tool selection and agent behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The content includes a runtime Deno-style import that fetches and executes remote code — "import { MCPServer, text } from 'https://esm.sh/mcp-use@latest/server'" — which, when used as shown, pulls and runs external code the skill would rely on at runtime.