What problem does it solve? AI agents often stall on trivial implementation choices, interrupting users with unnecessary questions, or conversely charge ahead into destructive actions. This Skill defines clear decision boundaries so agents make steady, reversible progress without blocking on the human, while still stopping for genuinely consequential choices. ## Core Features & Use Cases - Decision Triage: Distinguishes ordinary reversible execution choices from decisions involving intent, authority, scope, cost, or external impact that require user input. - Uninterrupted Workflow Completion: Finishes all phases of a sequenced task already in progress instead of pausing after each phase for confirmation. - Hard Safety Boundaries: Always stops for destructive or irreversible actions, security-sensitive changes, production data modifications, irreversible Git operations, and external communications. - Use Case: During a multi-file refactor, the agent chooses reasonable naming and structure from repository conventions, completes the work, and reports decisions and tradeoffs rather than asking about each file. ## Quick Start Ask the agent to complete a coding task end-to-end using the act-within-scope skill, making reasonable reversible decisions on its own and only stopping for irreversible or out-of-scope actions.
