{"skill":{"slug":"aana-task-scope-guardrail","displayName":"AANA Task Scope Guardrail Skill","summary":"Ensure the agent only performs requested tasks using relevant data, stops when complete, and avoids expanding beyond the user's original scope.","tags":{"latest":"1.0.0"},"stats":{"comments":0,"downloads":20,"installsAllTime":0,"installsCurrent":0,"stars":0,"versions":1},"createdAt":1777769046241,"updatedAt":1777769209393},"latestVersion":{"version":"1.0.0","createdAt":1777769046241,"changelog":"- Initial release of the AANA Task Scope Guardrail Skill.\n- Provides detailed instructions and principles to prevent agents from exceeding user-requested task boundaries or using unrelated data.\n- Defines clear scope categories, required scope checks, data relevance and expansion rules, and agent completion rules.\n- Includes an output pattern for scope-sensitive actions to ensure decisions are well-documented and within authorized boundaries.\n- Emphasizes minimal data use and requires explicit user approval before expanding task scope or accessing sensitive information.","license":"MIT-0"},"metadata":null,"owner":{"handle":"mindbomber","userId":"s177cynx168hg36ac5nrkb9wdn85z9dq","displayName":"mindbomber","image":"https://avatars.githubusercontent.com/u/42798111?v=4"},"moderation":null}