Add AgentGate — pre-execution authorization for agent tool use

[ElamOlame31]

What does this PR add / change?

Adds AgentGate to the Agent Security section — an open-source pre-execution authorization layer that intercepts AI agent actions before they execute and detects multi-step kill chain patterns across 24 hours.

Quality checklist

• The project is publicly accessible at the linked URL.
• At least one of:
  • 100+ GitHub stars, or
  • documented third-party adoption / production users, or
  • is an official artefact (vendor SDK, model card, government release)
• The description is one factual sentence — no marketing copy, no superlatives.
• Entry is placed alphabetically inside the appropriate section.

Independent traction (please show evidence)

1. Active discussion in microsoft/agent-governance-toolkit community (Discussion #276) referencing AgentGate's approach to pre-execution decision artifacts
2. PyPI package agentgate-pdp publicly available and installable
3. Python + TypeScript SDKs, LangGraph integration, comprehensive test suite — production-ready open source tool

Related submissions

• Also submitted to von-development/awesome-LangGraph (same day, full transparency)