The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill repeatedly instructs the user to install the Composio CLI using the command curl -fsSL https://composio.dev/install | bash. This pattern is a significant security risk as it downloads a script from a remote server and executes it directly in the shell without prior verification or auditing.
[COMMAND_EXECUTION]: The skill's primary functionality relies on executing a wide range of shell commands through the composio CLI. Examples include composio execute <TOOL_SLUG>, composio login, and composio search. If an AI agent incorporates untrusted user input into these commands without rigorous sanitization, it could lead to arbitrary command injection on the host system.
[EXTERNAL_DOWNLOADS]: The skill facilitates the installation of numerous external dependencies from public registries. This includes Python packages (e.g., composio, composio-openai-agents) and Node.js packages (e.g., @composio/core, @composio/vercel). While standard for such tools, these external dependencies introduce a supply chain risk surface.
[REMOTE_CODE_EXECUTION]: A URL included in the documentation for local testing purposes, https://abc123.ngrok.io/composio/triggers, has been flagged by security scanners as malicious (Phishing). Although it appears to be a placeholder/example for webhook testing, the reputation of the specific subdomain is compromised.

composio