Forge Blog
Tutorials, deep dives, and updates from the Forge team.
Forge is the runtime for agents that run next to services
The agent runtime market split in 2026. One half — desktop assistants and hyperscaler-managed agents — is well-served. The other half — agents that live inside your services, inside your network, on your infrastructure — is what Forge is built for.
The Claude Code Leak: What a .map File Reveals About the Future of Agent Security
Anthropic's Claude Code CLI source was exposed via a source map in npm. We analyze the OS-level architecture shift, expanded attack surface, and why compiled languages like Go prevent this class of vulnerability entirely.
OpenClaw vs Forge Architecture: Why Enterprise AI Agents Need Outbound-Only Networking
A visual comparison of OpenClaw's inbound gateway architecture and Forge's outbound-only model for enterprise AI agent deployment.
How to Run AI Agents Securely in Enterprise Environments (Without Inbound Tunnels)
Enterprise security teams need AI agents that don't require inbound tunnels. Learn how Forge's outbound-only architecture, egress allowlists, and structured audit logging solve enterprise agent security.
OpenClaw Demonstrated Agents Can Act. The Next Question Is: Can They Operate in an Enterprise?
AI agents are no longer hypothetical. But personal agents and enterprise agents are fundamentally different. Forge is the secure, scalable, enterprise-grade runtime for AI work.
Deploy a SKILL.md Agent in 5 Minutes
Walk through creating a Forge agent from scratch — write a SKILL.md, configure forge.yaml, and run it locally.
How Forge Computes Trust for Every Skill
A deep dive into Forge's trust model — how skills are scanned, analyzed, and assigned trust levels automatically.
Contributing Your First Skill to Forge
How to write, test, and submit a community skill to the Forge skill registry.