cc-sdd is a TypeScript-based tool that turns approved specs into long-running autonomous implementations across 8 coding agents, with a focus on a spec-driven SDLC workflow.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
cc-sdd is a long-running spec-driven implementation harness that coordinates agent skills across multiple AI coding agents (Claude Code, Codex, Cursor, Copilot, Windsurf, OpenCode, Gemini CLI, Antigravity). It provides a workflow from discovery to autonomous implementation with per-task independent review, using a set of 17 skills per agent and tooling to manage specs, design, tasks, and implementation.
How it works
The project exposes an SDLC workflow where discovery routes work into tasks and specs, followed by autonomous implementation with per-task reviewer passes and auto-debug. It supports multiple agent platforms via Skill variants and loads subagents on demand. It includes an integration point for long-running autonomous execution and per-task review loops, with a boundary-focused spec discipline and design notes propagating learnings through Implementation Notes.
Getting started
Quick start is shown as:
cd your-project
npx cc-sdd@latest
To pick another agent or language:
npx cc-sdd@latest --codex-skills --lang ja # Codex, Japanese
npx cc-sdd@latest --cursor-skills --lang zh-TW # Cursor IDE, Traditional Chinese
Supports 8 AI coding agents and 13 languages. See Supported Agents for full list. In use, run the entry point:
/kiro-discovery <idea>
Recent releases
Latest releases include:
- v3.0.2 (2026-04-13): Patch fixes for Codex spec-reviewer template description
- v3.0.1 (2026-04-11): Path safety hardening and mojibake fix
- v3.0.0 (2026-04-09): Skills mode and autonomous implementation across 8 platforms
- v2.1.1 (2026-02-01): Bug fix for OpenCode agent and security update
- v2.1.0 (2026-02-01): Added OpenCode support and updated recommended models
Traction
Stars: 3599 Forks: 274 Open issues: 26
License
MIT