Skip to content

madhukar04012/agentforge

v1.5.0Apache-2.0

AI Agent Development, Evaluation & Deployment Platform


AgentForge is an AI agent development platform for building, evaluating, and deploying production agents.


Turn your favorite coding assistant into an expert at building and deploying agents on Google Cloud.

AgentForge gives your coding agent the skills and commands to build, scale, govern, and optimize enterprise-grade agents — so you don't have to learn every CLI and service yourself.

Works seamlessly with: Antigravity CLI  •  Claude Code  •  Codex  •  and any other coding agent.

Get Started

Prerequisites: Python 3.11+, uv, and Node.js.

1. Install

uvx agentforge setup

Or just the skills — your coding agent will handle the rest

npx skills add Madhukar04012/AgentForge

2. Open your coding agent

Launch Antigravity CLI, Claude Code, Codex, or any coding agent you prefer.

3. Build your first agent

Ask your coding agent to build something — e.g. "Use AgentForge to build a caveman-style agent that compresses verbose text into terse, technical grunts"

See the full tutorial for a step-by-step walkthrough.

Browse the full documentation →


Agent Skills

SkillWhat your coding agent learns
agentforge-workflowDevelopment lifecycle, code preservation rules, model selection
agentforge-adk-codeADK Python API — agents, tools, orchestration, callbacks, state
agentforge-scaffoldProject scaffolding — create, enhance, upgrade
agentforge-evalEvaluation methodology — metrics, datasets, LLM-as-judge, adaptive rubrics
agentforge-deployDeployment — Agent Runtime, Cloud Run, GKE, CI/CD, secrets
agentforge-publishGemini Enterprise registration
agentforge-observabilityObservability — Cloud Trace, logging, third-party integrations

CLI Commands

The primary command is agentforge. The legacy agents-cli binary remains available as a compatibility alias.

CommandWhat it does
agentforge setupInstall CLI + skills to coding agents
agentforge create <name>Create a new agent project
agentforge eval runRun the agent over the eval dataset and grade the traces
agentforge deployDeploy to Google Cloud
agentforge publish gemini-enterpriseRegister with Gemini Enterprise

See all commands

CommandDescription
agentforge loginAuthenticate with Google Cloud or AI Studio
agentforge login --statusShow authentication status
Scaffold
agentforge create <name>Create a new agent project
agentforge scaffold enhanceAdd deployment or CI/CD to an existing project
agentforge scaffold upgradeUpgrade project to a newer AgentForge version
Develop
agentforge run "prompt"Run agent with a single prompt
agentforge installInstall project dependencies
agentforge lintRun code quality checks (Ruff)
Evaluate
agentforge eval runRun agent inference and grade the traces
agentforge eval generateRun agent inference over eval cases
agentforge eval gradeGrade generated traces against metrics
agentforge eval dataset synthesizeSynthesize multi-turn eval scenarios for your local agent
agentforge eval compareCompare two eval result files
agentforge eval analyzeCluster failure modes from grade results
agentforge eval metric listList available metrics
agentforge eval optimizeAuto-tune agent prompts using eval data
Deploy & Publish
agentforge deployDeploy to Google Cloud
agentforge publish gemini-enterpriseRegister with Gemini Enterprise
agentforge infra single-projectProvision single-project infrastructure
agentforge infra cicdSet up CI/CD pipeline + staging/prod infrastructure
Other
agentforge infoShow project config and CLI version
agentforge updateForce reinstall skills to all IDEs

How it works


Architecture

The Google Cloud agent stack that AgentForge builds on.

FAQ

Is this an alternative to Antigravity CLI, Claude Code, or Codex? No. AgentForge is a tool for coding agents, not a coding agent itself. It provides the CLI commands and skills that make your coding agent better at building, evaluating, and deploying ADK agents on Google Cloud.

How is this different from just using adk directly? ADK is an agent framework. AgentForge gives your coding agent the skills and tools to build, evaluate, and deploy ADK agents end-to-end.

Do I need Google Cloud? For local development (create, run, eval), no — you can use an AI Studio API key to run Gemini with ADK locally. For deployment and cloud features, yes.

Can I use this with an existing agent project? Yes. agentforge scaffold enhance adds deployment and CI/CD to existing projects.

Can I use AgentForge without a coding agent? Yes. The CLI works standalone — you can run agentforge scaffold, eval, deploy, and every other command directly from your terminal. The skills just make it easier for coding agents to do it for you.

How can I extend AgentForge with other skills? AgentForge skills cover the agent-building lifecycle (scaffold, ADK code patterns, evals, deploy, publish, observability). For adjacent concerns, you could install another skill suite alongside. For example, agent-skills covers general software-engineering workflows (ideation, spec gates, planning, code review), and google/skills covers Google Cloud foundations (BigQuery, Cloud Run, Firebase, GKE).

Feedback

We value your input — it helps us improve AgentForge for the community.

  • Bugs & feature requests: open an issue — 👍 the ones you want prioritized
  • Share what you built: we'd love to hear about your projects! Reach out at agents-cli@google.com to share your agent or provide feedback

Terms of Service

AgentForge leverages Google Cloud APIs. When you deploy agents, you'll be deploying resources in your own Google Cloud project and will be responsible for those resources. Please review the Google Cloud Service Terms for details.


Web Platform (Phase 1)

A web frontend for the same foundation is landing incrementally. The CLI is not changing — the web app is a thin FastAPI backend + a React/Vite SPA that wraps the existing eval flywheel, deploy targets, and ADK-backed agent runtime in a browser experience.

Phase 1 ships only the skeleton (settings, health, login/logout, a placeholder recruiter demo, and the Docker compose stack). Real domain endpoints (agents, workflows, executions, evaluations, deployments) arrive in later phases.

Bring up the local stack:

docker compose -f docker/docker-compose.yml up

Then:

  • http://localhost:8080 — web SPA
  • http://localhost:8000/api/v1/health — API liveness
  • http://localhost:8000/api/docs — OpenAPI / Swagger UI

See docs/architecture/architecture.md and docs/architecture/roadmap.md for the full design and phase tracker.