composio

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

• Third-party content exposure detected (high risk: 0.90). The skill's docs and workflow (e.g., the "Complete Auth + Trigger Flow Example" and triggers sections in auth-and-triggers.md / SKILL.md) explicitly subscribe to external trigger events like GITHUB_PULL_REQUEST_EVENT, GMAIL_NEW_EMAIL, and SLACK_NEW_MESSAGE and instruct passing that event data to the agent for processing, meaning the agent ingests untrusted, user-generated third‑party content (GitHub/Slack/Gmail, etc.) that can materially influence actions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill instructs clients to use the runtime MCP endpoint (session.mcp.url — the Composio MCP URL on composio.dev) which is fetched at runtime to deliver dynamic tool definitions/tool router that directly control the agent's available tools and behavior.