composio-mcp

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow uses Composio to connect AI agents to external apps (such as GitHub, Slack, Notion, Jira, and Gmail) via CLI or MCP, which can ingest outsider-authored comments, issues, and emails without the agent first selecting a specific item.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill instructs the agent to install third-party packages from npm via npm install -g composio-core @composio/cli, which constitutes an unpinned remote package dependency with arbitrary community provenance.