claude-allow-stop
Return to normal stop behavior. Use PROACTIVELY after completing all tasks in continuous work mode, when work is committed and tests pass, or when reaching natural stopping point. User can invoke with /claude-allow-stop.
When & Why to Use This Skill
The claude-allow-stop skill is a workflow control utility designed for Claude agents operating in continuous or autonomous modes. It provides a mechanism to restore standard stop behavior, ensuring the agent pauses for user input once specific milestones—such as task completion, successful test passes, or code commits—are achieved. This skill is essential for managing agentic autonomy and optimizing human-in-the-loop transitions in complex automation pipelines.
Use Cases
- Workflow Finalization: Proactively ending an autonomous session after the agent has completed a multi-step coding or research task to prevent unnecessary resource consumption.
- Quality Assurance Gates: Configuring the agent to stop and wait for review immediately after a test suite passes and changes are committed to the repository.
- Manual Override: Providing users with the /claude-allow-stop command to manually re-establish control over the agent's execution loop during long-running automated processes.
- Continuous Integration Sync: Ensuring the agent reaches a natural stopping point before handing off the environment to other automated scripts or human developers.
| name | claude-allow-stop |
|---|---|
| description | Return to normal stop behavior. Use PROACTIVELY after completing all tasks in continuous work mode, when work is committed and tests pass, or when reaching natural stopping point. User can invoke with /claude-allow-stop. |
| allowed-tools | Bash, Read |
Disable Stop Hook Automation
Allow normal completion and wait for user input.
.claude/skills/claude-allow-stop/scripts/allow-stop.sh
Or user invokes: /claude-allow-stop
Re-enable with /claude-do-not-stop [reason]