{"slug": "trajeckt-a-fail-closed-gateway-that-enforces-what-ai-agents-can-do-1-6ms", "title": "Trajeckt: a fail-closed gateway that enforces what AI agents can do (~1.6ms)", "summary": "Trajeckt, a new fail-closed gateway, enforces AI agent actions across entire sequences of tool calls rather than inspecting them one at a time, blocking violations that span multiple steps before irreversible actions fire. The system operates below the model at runtime with sub-2 millisecond latency, integrating as a configuration change for any stack using MCP or OpenAI-compatible tool calls. Trajeckt turns agent executions into a structured evidence layer for live feeds, policy state, replay history, and learning signals, enabling governance and observability across the full trajectory.", "body_md": "# A single tool call is\n\nrarely the problemThe sequence is.\n\nEvery agent guardrail today inspects actions one at a time — so the failures that only exist across steps walk right through. Trajeckt enforces commitments across the whole trajectory, at runtime, below the model, before any irreversible action fires.\n\n## Learn how your agents think\n\nTrajeckt turns executions into a structured evidence layer: live feeds, policy state, replay history, and the learning signals that shape better runs.\n\nFlip it. The plan doesn't change. The outcome does.\n\nGo ahead. It's a replay.\n\nThe injected step was off the declared plan, so it was blocked before it fired: the same tool, allowed in plan and stopped out of it.\n\n1.7msScripted from benchmark run-20260604-152135\n\n## One Enforcement Layer.\n\nEverything else Follows.\n\nTrajeckt turns executions into a structured evidence layer: live feeds, policy state, replay history, and the learning signals that shape better runs.\n\n## Integration.\n\nA gateway, not a library: if your stack speaks MCP or OpenAI-compatible tool calls, enforcement is one config change.\n\nAny MCP client. The highlighted line is the server URL — the only change.\n\n## Turn every trajectory into operating intelligence.\n\nStart with a control surface built for agent observability, governance, and investigation. Trajeckt makes it easier to see what happened, enforce what should happen, and improve what happens next.", "url": "https://wpnews.pro/news/trajeckt-a-fail-closed-gateway-that-enforces-what-ai-agents-can-do-1-6ms", "canonical_source": "https://traject.tamor.ai/", "published_at": "2026-06-12 13:39:49+00:00", "updated_at": "2026-06-12 13:49:59.475976+00:00", "lang": "en", "topics": ["ai-safety", "ai-agents", "ai-tools", "ai-infrastructure", "ai-products"], "entities": ["Trajeckt", "MCP"], "alternates": {"html": "https://wpnews.pro/news/trajeckt-a-fail-closed-gateway-that-enforces-what-ai-agents-can-do-1-6ms", "markdown": "https://wpnews.pro/news/trajeckt-a-fail-closed-gateway-that-enforces-what-ai-agents-can-do-1-6ms.md", "text": "https://wpnews.pro/news/trajeckt-a-fail-closed-gateway-that-enforces-what-ai-agents-can-do-1-6ms.txt", "jsonld": "https://wpnews.pro/news/trajeckt-a-fail-closed-gateway-that-enforces-what-ai-agents-can-do-1-6ms.jsonld"}}