iblai-agent-chat

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

• Third-party content exposure detected (high risk: 0.85). The required workflow wires the assistant to the hosted iblai-agent-chat MCP server, and at runtime the assistant sends user chat messages and receives the agent’s streamed replies (outsider-authored free text from the remote agent) into the LLM context via the MCP tool responses.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill configures and uses the hosted MCP server at https://asgi.data.iblai.app/mcp/agent-chat/ as the runtime transport for live chat (streamed responses, tools, and RAG), so the remote server directly controls conversational outputs/behavior at runtime.