Stats
Actions
Tags
Help us improve
Share bugs, ideas, or general feedback.
From ar
Resumes autorun workflow for uninterrupted, fully autonomous, non-interactive task execution with safety protocols and stop signals.
npx claudepluginhub ahundt/autorun --plugin arHow this command is triggered — by the user, by Claude, or both
Slash command
/ar:goThe summary Claude sees in its command listing — used to decide when to auto-load this command
# Resume Claude automatically when it stops - Autorun Workflow (short: /ar:go) $ARGUMENTS Your primary objective is to continue the **UNINTERRUPTED, FULLY AUTONOMOUS, NONINTERACTIVE, PATIENT, AND SAFE EXECUTION** of this /autorun workflow including your current tasks and goals. 1. **MANDATORY PROCESS TO CONTINUE EXECUTION:** Cautiously and deliberately continue working **carefully, patiently, concretely, and safely**, non-stop, autonomously, and non-interactively per your instructions and CLAUDE.md philosophy and definition of concrete. * **Permissions Granted:** You have full permiss...
/next-todoQueues a task to automatically start immediately after the current task finishes, without interrupting ongoing work or requiring confirmation.
/resumeResumes a previously-running Claude agent by ID, picking up its transcript and continuing the interrupted task.
/goOrchestrates goal-driven development: analyzes stated goal, creates/resumes track with spec/plan/metadata, runs autonomous evaluate-loop via orchestrator.
/goApprove a session plan and execute it wave by wave, or fast-track via Express Path when available. Also supports persisted audit-trail verification.
Share bugs, ideas, or general feedback.
$ARGUMENTS
Your primary objective is to continue the UNINTERRUPTED, FULLY AUTONOMOUS, NONINTERACTIVE, PATIENT, AND SAFE EXECUTION of this /autorun workflow including your current tasks and goals.
MANDATORY PROCESS TO CONTINUE EXECUTION: Cautiously and deliberately continue working carefully, patiently, concretely, and safely, non-stop, autonomously, and non-interactively per your instructions and CLAUDE.md philosophy and definition of concrete.
SIGNAL RULE TO STOP SYSTEM: The exact strings AUTORUN_STATE_PRESERVATION_EMERGENCY_STOP and AUTORUN_ALL_TASKS_COMPLETED_AND_VERIFIED_SUCCESSFULLY are SYSTEM STOP SIGNALS. You MUST NEVER output these strings unless the corresponding stop condition has been met. Accidental output will halt the entire system.
Safety Protocol (Risk Assessment & Mitigation): You MUST execute the full Preservation Sequence ONLY IF the current task involves a high-risk or irreversible destructive tool call (e.g., initial state modification, database interaction, or action following a prior failure).
CRITICAL ESCAPE TO STOP SYSTEM (Final Decision): Only if the risk is irreversible, catastrophic, or cannot be fully mitigated, you MUST initiate the Preservation Protocol by immediately outputting the following exact string to immediately halt all actions: AUTORUN_STATE_PRESERVATION_EMERGENCY_STOP
FINAL OUTPUT ON SUCCESS TO STOP SYSTEM (Final Decision): When all tasks and goals are all 100 percent complete and verified, remember you are often overconfident, so first double check the resources from which you receive or retrieve new tasks and think if anything was missed, you MUST conclude your entire final response with the exact string: AUTORUN_ALL_TASKS_COMPLETED_AND_VERIFIED_SUCCESSFULLY
FILE CREATION POLICY: ALLOW ALL: Full permission to create/modify files.