composio

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt instructs agents to obtain and then use CLI keys directly (e.g., retrieving a URL/key and running commands like composio login --key <cli_key>), which requires embedding secret values verbatim in generated commands/outputs.

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs agents to connect to and fetch user-generated content from external services (Gmail, Slack, GitHub, etc.) via the Composio CLI/SDK—see SKILL.md and AGENTS.md examples like composio execute, composio listen, and session.tools()/session.toolkits()—and those fetched third‑party messages/events are read and acted on by the agent, which can allow indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's setup instructs users to run a remote installer that is piped to the shell (curl -fsSL https://composio.dev/install | bash), which fetches and executes remote code and is a required runtime dependency for the CLI-based workflows, so this URL is a high-confidence risky external dependency.