iblai-api-agent-chat
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill provides a command to execute
claude mcp addfor registering the hosted MCP server. This is a standard and expected configuration step for the Claude Code tool. - [EXTERNAL_DOWNLOADS]: The skill configures a connection to a remote MCP service at
https://asgi.data.iblai.app/mcp/agent-chat/. This domain belongs to the skill's author/vendor, and the connection is necessary for the skill's stated functionality. - [CREDENTIALS_UNSAFE]: The instructions involve the use of an
IBLAI_API_KEY. The skill correctly advises sourcing this from an environment variable (.env) and explicitly warns against committing secrets to version control. - [PROMPT_INJECTION]: The skill enables a live chat interface with an external agent, which constitutes an indirect prompt injection surface. Chat responses from the remote service enter the local agent's context. No specific boundary markers or message sanitization techniques are documented, though this interaction is the primary intended purpose of the skill.
Audit Metadata