Configures and manages settings for AI coding agents including Claude Code and Codex CLI across multiple scopes.
The Settings Management skill provides a structured way to view and modify the configuration of AI coding agents, ensuring they operate within desired parameters. It handles complex tasks such as setting default LLM models, defining granular security permissions for tool use, managing sandbox environments, and configuring environment variables. By supporting a hierarchy of settings—from global user preferences to project-specific shared configs and local overrides—it enables developers to maintain a consistent and secure development environment while tailoring agent behavior to the specific needs of each repository.
主要功能
01Model selection and Extended Thinking mode configuration
0232 GitHub stars
03Multi-scope configuration management (User, Project, Local, and Managed)
04Granular permission control for tool access and file restrictions
05Support for both JSON (Claude Code) and TOML (Codex CLI) formats
06Sandbox and environment variable management
使用场景
01Enforcing project-wide security policies by restricting access to sensitive files like .env
02Switching between different Claude models or enabling thinking capabilities for complex tasks
03Configuring custom git attribution and commit hooks for AI-generated code
