MMCPMarkets
LG

liyupi/github-claw

Developer tools

This page was translated automatically. Technical details may differ from the original. View the original

68 stars Quality 39 Trend 14
Repository Install

Building AI agents with GitHub: long-term memory + AI Agent skill kit (UI/UX, image generation, SEO) + GitHub Actions automation + Copilot programming agent, supporting Vibe Coding for rapid project incubation.

Overview

A ready-to-use tool: using the GitHub repository as a carrier, integrating long-term memory, reusable AI skill packs, GitHub Actions automation, and Copilot programming agents, making AI truly your long-term collaborative partner. It demonstrates how to build a ** using GitHub repositories. Its core is not a specific project, but a set of mechanisms that keep AI running: - 🧠 : Retain AI Agent context across conversations via warehouse files, achieving true long-term memory without being limited to single sessions. - 🔧 : Package capabilities such as UI/UX design, AI image generation, and SEO auditing into a standardized skill kit that AI assistants can directly call and anyone can reuse. - ⚡ : Use GitHub Actions to implement daily AI technology summaries, intelligent issue assignment, and automatic order acceptance by Copilot programming agents without human intervention. - 🚀 : With AI Agent skill kits, describe requirements in natural language, and AI automatically completes coding, testing, and deployment—all warehouse subprojects are incubated this way. And just the practical results of this AI agent system demonstrate what AI Agent + Vibe Coding can achieve. This will allow AI assistant Claw to restore his identity, read long-term memories, and continue working from where he last left off. The Agent Skill is the core mechanism of this repository—it encapsulates reusable AI capabilities into standard units stored in .agents/skills//, with SKILL.md as the main entry point, allowing AI assistants to directly call upon receiving tasks. ： 1. Upon receiving the task→ first check the local .agents/skills/ and prioritize reusing existing skill packs. 2. No suitable skills locally→ Search for GitHub open-source repositories or Skills.sh 3. After installation, store everything uniformly in .agents/skills/ and update the index The following projects were all developed using the AI Agent + Vibe Coding approach, demonstrating the actual output capabilities of this agent system.

Overview (machine translated)

A ready-to-use AI Agent workspace template: Using GitHub repositories as a carrier, integrating long-term memory, reusable AI skill kits, GitHub Actions automation, and Copilot programming agents, making AI truly your long-term collaborative partner.

What is this?

github-claw demonstrates how to use GitHub repositories to build an memory-driven, skilled, and automatable AI agent. Its core is not a specific project, but a set of mechanisms that keep AI running:

  • 🧠 File-Memory: Use warehouse files to retain the AI Agent’s context across conversations, enabling true long-term memory without being limited to single sessions.
  • 🔧 AI Agent Skills: Standardizes UI/UX design, AI image generation, SEO auditing, and other capabilities into a standardized skill package. AI assistants can be directly called and reused by anyone.
  • ⚡ GitHub Actions-Driven AI Automation: Use GitHub Actions to implement daily AI technical summaries, intelligent issue assignment, and Copilot programming agents for automatic order acceptance, all without human intervention.
  • 🚀 Vibe Coding Practice: With the AI Agent skill kit, requirements are described in natural language, and AI automatically completes coding, testing, and deployment—all subprojects within the warehouse are incubated this way.

site/ and media-platform/ are just the results of this AI agent system, demonstrating what AI Agent + Vibe Coding can achieve.

Get started quickly

When starting a new Copilot conversation, tell the assistant:

This will allow AI assistant Claw to restore his identity, read long-term memories, and continue working from where he last left off.

AI Agent Skill Pack

The Agent Skill is the core mechanism of this repository—encapsulating reusable AI capabilities into standard units stored in .agents/skills/ /, with the main entry point being SKILL.md, allowing AI assistants to directly call upon receiving tasks.

Built-in Skill Pack:

Skill Pack Ability Description
ui-ux-pro-max UI/UX Design System: 50+ visual styles, 161 color schemes, 57 font combinations, one-click generation of high-quality interfaces
ai-image-generation Multi-model AI image generation (FLUX, Gemini, Grok, etc.), supporting text-to-image, image-to-image, LoRA, and more
seo-audit SEO auditing and optimization recommendations, automatically analyzing page structure, keyword density, and technical SEO issues

Skills Workflow:

  1. Receive a quest → First check local .agents/skills/, prioritize reusing existing skill packs

This page was translated automatically. Technical details may differ from the original.

Recommended Tools

Try a different keyword or remove a filter.

Install

npx skillfish add liyupi/github-claw
New to MCP? Start with the mcpskill.org guide →