mcp-cli

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

• Third-party content exposure detected (high risk: 0.90). This skill explicitly connects to arbitrary remote MCP servers and ingests their tools, resources, and prompts (see SKILL.md / README examples like "mcp-to-cli connect https://mcp.notion.com/mcp" and the src/cli.ts flows that call client.listTools(), resources.get, and prompts.get), so untrusted third‑party content is read and can influence subsequent tool calls and agent behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The CLI connects at runtime to user-supplied MCP server URLs (e.g., https://mcp.notion.com/mcp) and fetches remote tools/resources/prompts which are then injected into the client flow and directly control prompts/instructions, so the skill relies on external runtime content from that URL.