Configures and manages the defense-in-depth security layers for Claude Code's execution environment.
The cg-security-config skill provides a specialized command-line interface within Claude Code to manage Context Guard's multi-layered security system. It allows developers to dynamically toggle security scanners for prompt injection and data leakage, adjust enforcement modes between warning and blocking, and manage a granular allowlist for trusted tools. This skill is essential for users who need to maintain a secure development environment by auditing and controlling how Claude interacts with sensitive files and external inputs.
主な機能
01Real-time configuration auditing with formatted status tables
02Dynamic security mode switching between log, warn, and block settings
030 GitHub stars
04One-command reset to restore factory-default security configurations
05Granular toggling for injection, leakage, manipulation, and file guard scanners
06Interactive allowlist management for specific tool patterns and exceptions
ユースケース
01Hardening Claude's environment against prompt injection during untrusted code reviews
02Preventing accidental data leakage of sensitive environment variables or credentials
03Authorizing specific legacy scripts or internal tools via the security allowlist
