By devsforge
Orchestrate autonomous multi-task development loops with checkpoint/rollback, queue-based task management, and bidirectional sync to a Control Center — enabling iterative AI-driven coding sessions with progress tracking and resumable workflows.
Abort the current FORGE loop
Adopt WebUI tasks into spec/plan structure
Run all queued tasks until queue is empty (batch mode)
Create a manual checkpoint
Mark current FORGE task as complete
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Disclaimer: FORGE is an independent, community-developed open-source project. It is not affiliated with, endorsed by, or officially associated with Anthropic, PBC or any of its products (including Claude, Claude Code, or the Claude API). "Claude" and "Anthropic" are trademarks of Anthropic, PBC. This project is a third-party plugin built by the community to extend Claude Code functionality.
Specification-Driven AI Development Engine for Claude Code
FORGE transforms Claude Code into a powerful iterative development system that autonomously works on complex tasks through formal specifications, structured planning, and completion criteria verification.
git clone https://github.com/devsforge/Forge.git
cd Forge
pnpm install
pnpm build
# Add the marketplace
/plugin marketplace add devsforge/forge
# Install the plugin
/plugin install forge@devsforge
Or for local development, add to ~/.claude/settings.json:
{
"plugins": [
"/path/to/Forge/packages/plugin"
]
}
The plugin provides:
/forge:* slash commands (21 total)cd /path/to/Forge
pnpm dev:server
# Opens at http://127.0.0.1:3344
Or run directly:
node packages/server/dist/index.js
/forge:forge-register
This creates .forge.json and registers your project with Control Center.
/forge:forge-spec "Add user authentication with JWT tokens"
This creates a formal specification with:
/forge:forge-plan spec-001
Generates a detailed plan with:
/forge:forge-queue --plan plan-001
# Single task mode (interactive)
/forge:forge
# Batch mode (process entire queue autonomously)
/forge:forge-batch
FORGE will:
| Command | Description |
|---|---|
/forge:forge-spec "desc" | Create specification from description |
/forge:forge-plan SPEC_ID | Create implementation plan |
/forge:forge-queue --plan PLAN_ID | Queue tasks for execution |
/forge:forge | Start single task execution |
/forge:forge-batch | Process entire queue autonomously |
/forge:forge-adopt TASK_ID | Formalize WebUI tasks into spec workflow |
/forge:forge-request | Process pending WebUI requests |
/forge:forge-done | Mark current task as manually completed |
| Command | Description |
|---|---|
/forge:forge-pause | Pause the current loop |
/forge:forge-resume | Resume a paused loop |
/forge:forge-abort | Abort and exit |
| Command | Description |
|---|---|
/forge:forge-checkpoint | Create manual checkpoint |
/forge:forge-rollback | Rollback to checkpoint |
| Command | Description |
|---|---|
/forge:forge-register | Register with Control Center |
/forge:forge-link PROJECT_ID | Link to existing project |
/forge:forge-sync | Full bidirectional sync |
| Command | Description |
|---|---|
/forge:forge-status | Check current status |
/forge:forge-history | View iteration history |
/forge:forge-tasks | List tasks |
/forge:forge-help | Show help |
Access the web dashboard at http://127.0.0.1:3344
npx claudepluginhub devsforge/forge --plugin forgeDevsForge Enterprise Feature Flag Architect delivering comprehensive progressive rollout methodologies, A/B testing frameworks, canary deployment strategies, and experimentation platforms that transform risky releases into controlled, data-driven feature delivery systems
Use this agent when evaluating new development tools, frameworks, or services for the studio. This agent specializes in rapid tool assessment, comparative analysis, and making recommendations that align with the 6-day development cycle philosophy. Examples:\n\n<example>\nContext: Considering a new framework or library
Desktop App Dev subagent
DevsForge license compliance checker with dependency scanning, license compatibility validation, and report generation
Use this agent when conducting user research, analyzing user behavior, creating journey maps, or validating design decisions through testing. This agent specializes in understanding user needs, pain points, and behaviors to inform product decisions within rapid development cycles. Examples:\n\n<example>\nContext: Understanding user needs for a new feature
Turn Claude Code into a plan-execute-validate loop with parallel work, intelligent retry, and memory
Long Task Harness for AI agents - task/feature-driven development with external memory
oh-my-zsh for Claude Code — 11 agents, 33 commands, 24 skills, 15 hooks + 9 examples (21 events), 9 rules, 4 MCP (minimal: playwright, context7, jina-reader, [email protected]). One-line install: curl -fsSL https://raw.githubusercontent.com/sangrokjung/claude-forge/main/install.sh | bash
Continuous self-referential AI loops for interactive iterative development, implementing the Ralph Wiggum technique. Run Claude in a while-true loop with the same prompt until task completion.
Feature development with code-architect/explorer/reviewer agents, CLAUDE.md audit and session learnings, and Agent Skills creation with eval benchmarking from Anthropic.
Production-grade engineering skills for AI coding agents — covering the full software development lifecycle from spec to ship.