Getting Started
eforge is an agentic build system that turns specifications into code. You describe what you want to build; eforge plans, implements, reviews, and validates it autonomously using a multi-stage pipeline across isolated worktrees.
Prerequisites
- Node.js 22+
- One of: Pi, Claude Code, or an npm-capable shell
- An LLM credential for the runtime you choose: a provider-specific API key or OAuth token for the recommended
piharness, or an Anthropic API key for the supported secondaryclaude-sdkharness
Install
Pi package (recommended)
Start with Pi, the recommended eforge execution harness for new users, if you want provider-flexible, local, inspectable agent orchestration.
pi install npm:@eforge-build/pi-eforge
/eforge:initAdd -l to write to project settings (.pi/settings.json) instead of your global Pi settings:
pi install -l npm:@eforge-build/pi-eforgeClaude Code plugin
Use the Claude Code plugin if Claude Code is already your daily environment. Claude Code can host the workflow while your active profile executes builds through the recommended Pi harness.
Run these three commands inside Claude Code:
/plugin marketplace add eforge-build/eforge
/plugin install eforge@eforge
/eforge:init
The /eforge:init command creates eforge/config.yaml with sensible defaults and adds .eforge/ to your .gitignore. It walks you through a Quick setup (one harness/provider with suggested tier models, including an optional separate implementation model) or a Mix-and-match flow (different harness, provider, or model per tier). Choose Pi for the recommended provider-flexible path; claude-sdk remains available as a supported Anthropic-specific secondary path. Starting June 15, 2026, Anthropic says Claude Agent SDK and claude -p usage no longer count toward Claude plan limits; eligible plans may receive a separate monthly Agent SDK credit, usage beyond that credit is billed at standard API rates when extra usage is enabled, otherwise requests stop, and API-key users remain pay-as-you-go. See https://support.claude.com/en/articles/15036540-use-the-claude-agent-sdk-with-your-claude-plan.
Standalone CLI
npx @eforge-build/eforge build "Add rate limiting to the API"Or install globally: npm install -g @eforge-build/eforge
For standalone use, run /eforge:init in Claude Code or Pi first to create eforge/config.yaml and an agent runtime profile.
Your First Build
Once eforge is installed and initialized, start a build from Claude Code or Pi:
/eforge:plan
The /eforge:plan skill guides a structured planning conversation - exploring scope, architecture, and risks - before handing off to the build pipeline. When you are ready to build:
/eforge:build
Or enqueue directly with a prompt:
/eforge:build Add a dark mode toggle to the settings page
The daemon picks up the queued plan and runs the full pipeline in the background. A web monitor at http://localhost:<port> (port deterministically assigned per project in the 4567-4667 range) tracks progress, cost, and token usage in real time.
From the standalone CLI:
eforge build "Add a dark mode toggle to the settings page"
eforge build plans/my-feature-prd.mdWhat Happens Next
- Formatting - eforge normalizes your input into a structured PRD.
- Planning - A planner agent assesses complexity and selects a workflow profile (Errand, Excursion, or Expedition), then writes a detailed plan or set of plans.
- Building - Builder agents implement each plan in isolated git worktrees, in parallel where the dependency graph allows.
- Review - Blind reviewers evaluate each plan's output without builder context. A fixer applies suggestions; an evaluator accepts only strict improvements.
- Merge - Completed plans merge back to your branch in topological order.
- Validation - Post-merge validation runs your configured commands. On failure, a validation-fixer agent attempts repairs.
Where to Look Next
- Concepts - How the pipeline works, what blind review means, and what harnesses do
- Configuration - The most important config options and how to tune them
- Profiles - Create and switch agent runtime profiles that control harness, model, and effort
- Playbooks - Build reusable workflow templates for recurring work
- Integrations - How to use eforge from Claude Code, Pi, the CLI, and external issue trackers
- Troubleshooting - Daemon startup, failed builds, and common error remedies
- Glossary - Definitions for eforge-specific terms such as profiles, worktrees, and playbooks
- CLI Reference - All CLI commands and flags
- Configuration Reference - Full
eforge/config.yamlschema