AG2 v1.1.0 (2026-09-24) adds TypeSafe Jev decision-only agents via TypeSafeConfig + response_schema, and strips reserved ag:/a2a: context keys on A2A/AG-UI/A2UI/NLIP to block remote pre-approval of gated tools.

Key Takeaways

  • ✓Decision agents: TypeSafeConfig + response_schema; bool/Enum/IntEnum as question shapes; probs on ModelMessage.metadata
  • ✓Jev rejects tools and streaming; invalid schemas raise locally before the wire
  • ✓Install: pip install "ag2[typesafe]"; see TypeSafe Jev blog + docs.ag2.ai
  • ✓Security: strip reserved ag:/a2a: prefixes both ways on A2A/AG-UI/A2UI/NLIP
  • ✓Also: MessageEnqueued, MCP required-arg validation + JSON-RPC codes, event retarget fix
🔗

Project Links & Resources

Direct Access
Direct access to official project resources and documentation
🐙
AG2 v1.1.0 Release
github.com/ag2ai/ag2/releases/tag/v1.1.0
↗
🌐
PyPI ag2 1.1.0
pypi.org/project/ag2/1.1.0/
↗
📝
TypeSafe Jev blog
typesafe.ai/blog/introducing-system-one-models-and-jev
↗
📖
AG2 Docs
docs.ag2.ai/
↗
🌐
AG2 homepage
ag2.ai/
↗
🔬

In-Depth Technical Analysis

Background

Routing/handoff in multi-agent systems was often fragile prompts or fake tool calls. Mixing control-plane keys into conversation variables also let remotes pre-approve gated tools. AG2 1.1.0 adds TypeSafe Jev decision agents and strips reserved context across transports.

Architecture

TypeSafe Jev answers typed questions with distributions (no text). response_schema is the question (bool/Enum/IntEnum); else raise locally. Probs/confidence on ModelMessage.metadata. Tools and streaming unsupported. Reserved ag:/a2a: prefixes stripped both ways on A2A/AG-UI/A2UI/NLIP. Also: MessageEnqueued, MCP required-arg validation, JSON-RPC error codes.

Benchmarks

No public SWE scores in the release notes; verify via changelog schema checks and the reserved-key security fix.

Get started

pip install -U "ag2[typesafe]"; read the v1.1.0 notes and docs.