Open-Source Tools for AI‑Native Development
Self-hosted security gateway and workflow platform for AI developers. Protect agents, automate coding workflows, and ship faster — all open source.
Forge — AI Workflow Platform
Forge is a self-hosted web platform for Claude Code that lets you vibe code from anywhere. Run AI coding tasks in background with a persistent tmux-backed terminal, manage multiple projects from a single dashboard, and access everything remotely via Cloudflare Tunnel or Telegram Bot. Define multi-step pipelines in YAML, browse and install skills from the marketplace, track token costs per task, and choose models (Sonnet, Opus, Haiku) for each job. No API keys needed — uses your existing Claude Code subscription.
# MORE TOOLS
Bastion — AI Agent Security Gateway
A self-hosted local proxy that sits between your AI agents and LLM providers. It scans every prompt and response for data leaks (20 patterns including AWS keys, GitHub tokens, credit cards), detects prompt injection in real time with a local ML classifier, blocks dangerous tool calls before they execute (26 rules covering rm -rf, curl|bash, force push, .env reads), and logs everything with AES-256-GCM encryption.
Accord — Contract-First Multi-Service Orchestration
Coordinate AI agents across microservices using API contracts as the source of truth. A central Hub dispatches work to services maintained by AI agents, human developers, or a mix of both. Built on the A2A protocol (Linux Foundation) for standardized agent-to-agent communication. Designed for real-world projects like re-architecting enterprise platforms with 6+ microservices.
Agent Security Scanner — Free Online Tool
Test if your AI agent is vulnerable to indirect prompt injection in 60 seconds. We generate a honeypot URL disguised as SDK documentation, your agent reads it, and we show you exactly which attacks it fell for — stealth canary tokens, injected marketing claims, content suppression, command execution, and more. No install required.