agent-estate

Intercept Claude Code stop events and re-inject prompts for persistent autonomous loops.

4|Updated Feb 26, 2026
One-click install
npx skills add https://github.com/MercuriusDream/agent-estate --skill agent-estate
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-estate
Source: https://github.com/MercuriusDream/agent-estate/tree/main
Command: npx skills add https://github.com/MercuriusDream/agent-estate --skill agent-estate

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Agent Estate addresses the need for sustained, automated AI work loops by ensuring Claude Code continues to operate across cycles without forgetting prompts or context.

Core Features & Use Cases

  • Persistent ledger across sessions enabling cross-context memory.
  • Stop-hook based perpetual operation that re-injects prompts after each cycle.
  • Flexible start options via start commands and a done mode for auto-stop when tasks are complete.

Quick Start

Invoke the setup-estate.sh script to activate the perpetual loop.

Frequently Asked Questions about agent-estate

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I keep a Claude Code agent running continuously across multiple cycles?

A perpetual autonomous agent loop intercepts Claude Code stop events and re-injects prompts to sustain continuous operation. This mechanism prevents the agent from halting after completing a single cycle.

How does a persistent ledger maintain state for long-running autonomous agents?

A persistent ledger maintains state for autonomous agents by recording progress and context across sessions. This cross-session memory ensures the agent retains prompts and previous work without losing continuity.

How do I set up a perpetual loop for Claude Code automation?

To set up a perpetual loop for Claude Code automation, invoke the setup-estate.sh script. This activates the stop-hook based operation and initializes the persistent ledger for cross-context memory.

Can I auto-stop a Claude Code autonomous agent when a task is complete?

Yes, you can auto-stop a Claude Code autonomous agent by enabling a done mode. This mode automatically stops the perpetual re-injection of prompts when the agent determines that the assigned tasks are complete.

Does Claude Code support hooks for state management and autonomous workflows?

Claude Code supports hooks for state management by allowing plugins to intercept stop events and maintain a persistent ledger. This enables autonomous workflows where state and memory persist across long-running cycles.

When should I not use a stop-hook based perpetual operation for AI workflows?

You should not use a stop-hook based perpetual operation for simple, short-lived AI tasks that require immediate termination. It is designed for long-running code-generation, testing, and automation workflows requiring persistent state.