How does this skill handle Claude's session limits?

The skill uses an auto-continuation loop and a headless mode pattern to automatically start new sessions, passing the current state and task list forward until the project is finished.

Where is the progress of my autonomous tasks saved?

All progress, including task lists and session notes, is stored locally in your project root under a directory named .autonomous/ .

Can I stop an autonomous task and resume it later?

Yes. You can interrupt the process at any time, and by using the 'autonomous' trigger phrase later, the skill will detect the existing directory and offer to continue from where it left off.

Does this skill require manual supervision?

While designed for autonomous execution, it provides clear progress reports after every session, allowing you to monitor the agent's logic and intervene if necessary.

Autonomous Task Executor

Name: Autonomous Task Executor
Author: tianzecn

bytianzecn

0•

生产力与工作流

Executes complex, long-running tasks by decomposing them into manageable steps and automating multi-session execution.

The Autonomous Task Executor is a powerful skill designed to handle large-scale development projects that exceed the limits of a single AI session. By employing a dual-agent architecture—consisting of an Initializer for planning and an Executor for implementation—it breaks down high-level requests into granular task lists. The skill manages state persistence within a dedicated local directory, allowing Claude to track progress, handle session transitions automatically, and work through complex workflows like building full-stack features or performing major refactors without manual intervention between every step.

主要功能

01Automatic task decomposition into structured markdown task lists

02Headless mode support with configurable auto-continuation loops

030 GitHub stars

04Persistent state management via local .autonomous directory

05Dual-agent execution pattern (Initializer and Executor modes)

06Multi-task isolation for running parallel autonomous workflows

使用场景

01Executing large-scale codebase refactoring across multiple modules and files

02Building complete REST APIs or microservices from scratch through automated sessions

03Managing long-running maintenance tasks like documentation generation or comprehensive testing

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tianzecn/myclaudecode autonomous-skill

For use in Claude.ai and ChatGPT

主要功能

01Automatic task decomposition into structured markdown task lists

02Headless mode support with configurable auto-continuation loops

030 GitHub stars

04Persistent state management via local .autonomous directory

05Dual-agent execution pattern (Initializer and Executor modes)

06Multi-task isolation for running parallel autonomous workflows

使用场景

01Executing large-scale codebase refactoring across multiple modules and files

02Building complete REST APIs or microservices from scratch through automated sessions

03Managing long-running maintenance tasks like documentation generation or comprehensive testing

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tianzecn/myclaudecode autonomous-skill

For use in Claude.ai and ChatGPT