setup-skill-packs
Fail
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADS
Full Analysis
- [REMOTE_CODE_EXECUTION]: Fetches and executes the Bun installation script from bun.sh's official URL using a pipe to the shell.
- [REMOTE_CODE_EXECUTION]: Clones third-party GitHub repositories and executes their contained setup scripts and build commands, such as bun run build and npm install.
- [COMMAND_EXECUTION]: Performs extensive shell operations to configure the environment, including directory manipulation, symbolic link creation, and environment variable modification.
- [DATA_EXFILTRATION]: Reads the ~/.claude/settings.json configuration file, which may contain sensitive environment details or existing tool configurations.
- [CREDENTIALS_UNSAFE]: Instructs the user to provide and store API keys for external services like GitHub, Vercel, and Supabase during the MCP server configuration phase.
- [EXTERNAL_DOWNLOADS]: Downloads over 100 external skills and dozens of agents and commands from third-party repositories on GitHub.
- [COMMAND_EXECUTION]: Installs persistent Hooks and Agents into the core Claude Code configuration directory, which can intercept tool usage and modify agent behavior across all future sessions.
Recommendations
- HIGH: Downloads and executes remote code from: https://bun.sh/install - DO NOT USE without thorough review
- AI detected serious security threats
Audit Metadata