eforge
An open source agentic build system for developers who want to stay close to the work. Describe what you want built - a prompt, a markdown file, a full PRD - and hand it off. eforge plans the implementation, builds in isolated worktrees, reviews the code blind, and validates the result. The build runs in the background. Your careful planning gets executed faithfully.
The name: E from the Expedition-Excursion-Errand methodology + forge - shaping code from plans.
Status: This is a young project moving fast. Used daily to build real features (including itself), but expect rough edges - bugs are likely, change is expected, and YMMV. Source is public so you can read, learn from, and fork it. Not accepting issues or PRs at this time.
What is an Agentic Build System?
Traditional build systems transform source code into artifacts. An agentic build system transforms specifications into source code - then verifies its own output.
The key insight: a single AI agent writing and reviewing its own code will almost always approve it. Quality requires separation of concerns - distinct agents for planning, building, reviewing, and evaluating.
An agentic build system applies build-system thinking to this multi-agent pipeline:
- Spec-driven - Input is a requirement, not a code edit. The system decides how to implement it.
- Multi-stage pipeline - Planning, implementation, review, and validation are separate stages with separate agents, not one conversation.
- Blind review - The reviewer operates without builder context (see below).
- Dependency-aware orchestration - Large work decomposes into modules with a dependency graph. Plans build in parallel across isolated git worktrees, merging in topological order.
- Adaptive complexity - The system assesses scope and selects the right workflow: a one-file fix doesn't need architecture review, and a cross-cutting refactor shouldn't skip it.
Typical Use
Plan a feature interactively in Claude Code or Pi, then hand it off with /eforge:build. The extension enqueues the input and a daemon picks it up - planning, building, reviewing, and validating autonomously. A web monitor (default localhost:4567) tracks progress, cost, and token usage in real time.
eforge also runs standalone. By default, eforge build enqueues and a daemon processes it. Use --foreground to run in the current process instead.
How It Works
Formatting and enqueue - Whatever you hand eforge - a prompt, rough notes, a session plan, a detailed PRD - gets normalized into a structured PRD and committed to a queue directory on the current branch. The daemon watches this queue and picks up new PRDs to build.
Workflow profiles - The planner assesses complexity and selects a profile:
- Errand - Small, self-contained changes. Passthrough compile, fast build.
- Excursion - Multi-file features. Planner writes a plan, blind review cycle, then build.
- Expedition - Large cross-cutting work. Architecture doc, module decomposition, cohesion review across plans, parallel builds in dependency order.
Blind review - Every build gets reviewed by a separate agent with no builder context. Separating generation from evaluation dramatically improves quality - solo agents tend to approve their own work regardless. A fixer applies suggestions, then an evaluator accepts strict improvements while rejecting intent changes. The goal is fidelity to the plan - minimizing drift and slop so the code that lands is what was specified, not a reinterpretation.
Parallel orchestration - Each plan builds in an isolated git worktree. Expeditions run multiple plans in parallel, merging in topological dependency order. Post-merge validation runs with auto-fix.
Queue and merge - Completed builds merge back to the base branch as merge commits via --no-ff, preserving the full branch history while keeping first-parent history clean. When the next build starts from the queue, the planner re-evaluates against the current codebase - so plans adapt to changes that landed since they were enqueued.
