{"skill":{"slug":"claw-guardirails-skill","displayName":"ClawGuardrails","summary":"Enforces guarded execution with safe_exec, safe_send, and safe_action. Use when a task may run shell commands, send channel messages, or call external APIs/a...","tags":{"latest":"0.1.0"},"stats":{"comments":0,"downloads":364,"installsAllTime":0,"installsCurrent":0,"stars":0,"versions":1},"createdAt":1772132180615,"updatedAt":1777525420602},"latestVersion":{"version":"0.1.0","createdAt":1772132180615,"changelog":"- Initial release of guardrails-safe-tools skill.\n- Enforces safe execution for shell commands, outbound messages, and external actions via `safe_exec`, `safe_send`, and `safe_action`.\n- Provides strict rules for using only guarded tools instead of direct execution/messaging.\n- Ensures workflows stop for approval when required; explains decision handling for `allow`, `deny`, and `require_approval`.\n- Documents best practices for input hygiene and choosing safer defaults.","license":null},"metadata":null,"owner":{"handle":"eveiljuice","userId":"publishers:eveiljuice","displayName":"Timo","image":"https://avatars.githubusercontent.com/u/152583228?v=4"},"moderation":null}