HomeProjectsArticlesPK BoardAbout
Explore Projects

GSD-2

Stale
GitHub TypeScript MIT

Description

GSD-2 is a powerful meta-prompting, context engineering, and spec-driven development system that enables agents to work autonomously for long periods without losing track of the big picture.

Key Features

  • Meta-prompting system for long-running autonomous agent tasks
  • Context engineering to maintain coherence across extended sessions
  • Spec-driven development workflow for structured agent outputs
  • Agents stay on track with big-picture goals without drifting
  • TypeScript-based implementation for type-safe agent orchestration
  • Lightweight and focused on agent autonomy and task decomposition

Use Cases

💡 Long-running codebase refactoring tasks that require sustained context
💡 Multi-step feature implementation following detailed specifications
💡 Autonomous bug investigation and fixing across large codebases
💡 Complex project scaffolding with multiple interdependent components
💡 Research-driven development requiring persistent context across sessions

Strengths & Limitations

✅ Strengths

  • • High community interest (7.8k stars)
  • • Permissive open-source license (MIT)
  • • Responsive to issues, low backlog

Tags

workflow agent typescript coding automation

Categories

🔄 Workflow

Quick Start

Note: This repository has been archived. The project continues as GSD Pi at https://github.com/open-gsd/gsd-pi. Visit the new repository for current source code, issues, and releases.
Visit GitHub

Related Projects

AI Website Cloner

33.7k · JavaScript
Active A

Clone any website with one command using AI coding agents. Built with Next.js, React, and shadcn-ui with web scraping and automated code generation.

typescriptagentcoding +3
  • · Clone any website with one command — auto-screenshots, extracts design tokens, generates component specs, and builds a Next.js codebase
  • · Multi-phase pipeline: Recon → Foundation → Component Specs → Parallel Build → Assembly & QA for structured engineering workflow
  • · Supports 13 AI coding agents including Claude Code, Codex CLI, Cursor, Copilot, and more
View full details → View on GitHub

Planning with Files

26.6k · Shell
Active A+

Claude Code skill implementing Manus-style persistent markdown planning — the structured workflow pattern for agent task management.

agentworkflowcoding +2
  • · Manus-style markdown planning — decompose agent tasks into structured plans in persistent markdown files
  • · Cross-session task persistence — plan files survive across sessions, agents resume anytime
  • · Plan file locking mechanism — atomic writes and flock mutex to prevent concurrency conflicts
View full details → View on GitHub

Budibase

28.3k · TypeScript
Active A

An open-source low-code platform for building AI agents, automations and business applications, model agnostic with drag-and-drop visual development.

workflowtypescriptagent +2
  • · AI Agent builder — Built-in agent engine that understands natural language requests and executes workflows
  • · Drag-and-drop low-code development — Visual business app builder requiring no code
  • · Multi-source integration — Connect databases, AI models, and business apps for cross-system agent actions
View full details → View on GitHub

Better Chatbot

1.2k · TypeScript
Active A

A better chatbot platform powered by Agent, MCP, and Workflows. Supports multi-model integration, visual workflow orchestration, and low-code configuration.

mcpagentchatbot +5
  • · Multi-model support: integrates OpenAI, Anthropic, Google, xAI, Ollama, and all major LLMs
  • · MCP protocol driven: invoke browser automation, web search, code execution tools via @mention
  • · Visual workflow editor: drag-and-drop to build AI workflows and publish them as chat tools
View on GitHub