Manages automated guard scripts that prevent Claude from repeating past coding mistakes across different projects.
Respect Guards is a specialized management system for persistent AI memory, allowing developers to curate and maintain custom guardrails that prevent recurring errors. By organizing scripts into global and local scopes, the skill enables Claude to automatically detect risky commands or patterns—such as Friday deployments or incorrect configuration paths—and warn the user before execution. It provides a robust framework to list, fix, test, and delete guards, effectively turning hard-learned lessons into executable safety checks that improve reliability across sessions.
主な機能
01Multi-scope management for global and project-specific guard scripts
020 GitHub stars
03Integrated testing environment to validate guard logic against sample inputs
04Automated guard discovery with metadata extraction and status reporting
05Interactive troubleshooting for fixing misfiring or outdated guards
06Seamless synchronization with the /oops feedback loop system
ユースケース
01Creating persistent warnings for deprecated terminal commands or insecure patterns
02Preventing accidental production deployments during restricted time windows
03Enforcing project-specific architectural constraints and directory structures
