skills.rest
Earths-Reckoning
Earths-ReckoningCOMMUNITY
Explore all skills by this author →
GitHub

autonomous-tasks

Load and execute Markdown task files with persistent progress and verification.

Updated Aug 23, 2026
Search Tags:#task management#workflow automation#software development#ai agent#persistence#autonomous agent#context window
One-click install
npx skills add https://github.com/Earths-Reckoning/Cheeseburg124 --skill autonomous-tasks
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: autonomous-tasks
Source: https://github.com/Earths-Reckoning/Cheeseburg124/tree/main/.claude/skills/autonomous-tasks
Command: npx skills add https://github.com/Earths-Reckoning/Cheeseburg124 --skill autonomous-tasks

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires npm, anthropic, Pillow.

What problem does it solve?

This Skill empowers Claude to work through multiple tasks systematically and autonomously, even across context window refreshes. It ensures persistence, verified completion, and prevents the AI from stopping at the first error or claiming completion without proof.

Core Features & Use Cases

  • Task Queue Management: Load, sort, and execute tasks defined in Markdown files, prioritizing high-impact work.
  • Persistent Execution: Automatically save and restore progress across context window limits, ensuring continuity for long-running projects.
  • Verified Completion: Integrates with physical verification to ensure all implemented features actually work as intended, providing concrete evidence.
  • Use Case: Provide Claude with a list of 5 UI improvements (e.g., "add dark mode," "fix button alignment"). This skill will guide Claude to implement each, verify it visually, save progress, and continue until all are done, even if the conversation is long.

Quick Start

  1. Create task files in the '.claude/tasks/' directory (e.g., '001-cost-dashboard-minimize.md').
  2. Instruct Claude: "Load and execute all tasks in the '.claude/tasks/' directory autonomously. Ensure each task is physically verified before marking it complete."