Back to feed
Dev.to
Dev.to
5/12/2026
Every AI toolchain is inventing its own safety layer. We shipped one that works across all of them.

Every AI toolchain is inventing its own safety layer. We shipped one that works across all of them.

Short summary

JamJet solves fragmented action control across AI runtimes: Claude Code (PreToolUse hooks), OpenAI Agents SDK (guardrails), and MCP servers (proxy shims) now share one policy.yaml and unified audit log. The portable control plane replaces per-platform security configs with glob-based rules. Phase 2 ships integration packages for immediate adoption.

  • •Single policy.yaml controls Claude Code, OpenAI Agents SDK, and MCP runtimes
  • •Unified audit trail eliminates fragmented security configurations
  • •Five new packages shipped for immediate integration and adoption

Generated with AI, which can make mistakes.

#ai-agents#ai-tools#open-source#product-launch
Read full article at Dev.to

Is this a good recommendation for you?

Explore more

My AI agent wiped my database twice. So I built a command firewall.
Dev.to
Three Layers of Tool Call Hardening for AI Agents
Dev.to
The Missing Layer in Agent Security
Dev.to
Have you ever told an AI 'never do this' and watched it do it anyway?
Dev.to