tikhub-mcp

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

• Third-party content exposure detected (high risk: 0.75). The required workflow uses mcp-remote to fetch and proxy hosted MCP content from https://mcp.tikhub.io/{platform}/mcp at runtime, which is outsider-authored free text/tool descriptions that can be ingested into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes npx mcp-remote at runtime pointed at https://mcp.tikhub.io/{platform}/mcp (see the .mcp.json args), which fetches remote MCP tool definitions that directly determine the agent's tools/instructions and thus controls runtime behavior.