Neura Market
prompt logo

prompt

Free

Build AI agents that gracefully handle user interruptions and plan changes

FreeFree tier
Inputs: textOutputs: text
Type
Open Source
Visit Website GitHub (40 stars)

About prompt

The Interruptible Agent Planner is a prompt designed for AI agents that need to handle multi-step tasks while remaining responsive to user interruptions, priority changes, and partial cancellations. It enforces a structured approach to maintain a live task state (completed, in-progress, pending, blocked, irreversible actions), treat interruptions as first-class events, protect against stale intent, and prefer reversible progress. The output format includes sections for current objective, latest user change, state snapshot, work to stop, work to preserve, revised plan, confirmation needed, and irreversible risks. This prompt is ideal for long-running agent applications where the user may revise scope at any time, ensuring the agent does not continue executing an old plan after a change of direction. It is published as open source on GitHub under the ai-boost/awesome-prompts repository.

Key Features

Maintains a live task state: completed, in-progress, pending, blocked, and irreversible actions
Treats interruptions as first-class events, detecting changes in goal, constraints, or priority
Protects against stale intent by stopping non-essential work and preserving useful state
Prefers reversible progress by checkpointing before side effects and separating analysis from commitment
Outputs structured sections: Current Objective, Latest User Change, State Snapshot, Work to Stop, Work Safe to Preserve, Revised Plan, Confirmation Needed, Irreversible Risks
Explicitly distinguishes reversible from irreversible state and shows delta from previous plan

Pros & Cons

Pros
  • Designed specifically for handling user interruptions, making it suitable for dynamic environments
  • Forces explicit state tracking, reducing risk of executing stale plans
  • Allows agents to distinguish between reversible and irreversible progress
  • Open source and free to use, modify, and integrate
  • Based on research and best practices from arXiv, Anthropic, and OpenAI
Cons
  • Requires manual integration into an AI system; it is a prompt template, not a ready-to-use tool
  • Assumes the agent can follow a strict output format, which may not suit all frameworks
  • No built-in memory or task scheduling; external infrastructure needed for state persistence
  • Limited community support beyond the GitHub repository

Best For

Long-running AI agent tasks where users may change their mind mid-executionMulti-step workflows requiring dynamic re-planning upon user feedbackInteractive AI assistants that need to handle partial cancellations and priority shiftsApplications where irreversible actions must be confirmed before execution

FAQ

What is the Interruptible Agent Planner?
It is a prompt template for AI agents to carry out multi-step work while remaining ready to absorb user interruptions, priority changes, and partial cancellations without losing control of the task.
How does it handle interruptions?
It treats interruptions as first-class events: it detects whether the user changed goal, constraints, or priority, stops non-essential work quickly, preserves useful state, and re-plans from the updated objective.
What output format does it use?
It requires exactly eight sections: Current Objective, Latest User Change, State Snapshot, Work to Stop Immediately, Work Safe to Preserve, Revised Plan, Confirmation Needed, and Irreversible Risks.
Is this tool free?
Yes, it is an open-source prompt file hosted on GitHub under the ai-boost/awesome-prompts repository, free to use and modify.