Arbitype v0.7.0 — Professionalized typed decision tools for AI agents

Arbitype v0.7.0

Arbitype is an MCP-native typed decision layer for AI agents, powered by TypeSafe Jev.

Highlights

• Safe MCP host setup for Codex, Claude, Cursor and VS Code
• Interactive plan → diff → confirmation workflow
• Clear USE WHEN / DO NOT USE WHEN tool boundaries
• 120-case public tool-selection benchmark framework
• Recorded 96.67% Jev-mediated tool-selection result
• Corrected per-case decision stability methodology
• Real support-routing, PR-verification, release-review and agent-routing examples
• GitHub CodeQL and protected-main governance
• Full-SHA-pinned GitHub Actions and provenance attestations
• Zero runtime dependencies retained
• MCP 2026-07-28 interoperability retained

Benchmark scope

The 96.67% result is a Jev-mediated tool-selection benchmark over
Arbitype's advertised MCP tool catalog.

It is not a Codex, Claude, or Cursor host benchmark and is not a
general model-performance guarantee.

Compatibility and safety

Legacy imports and CLI aliases remain available through the Arbitype distribution.
Probabilities and confidence are model signals, not proof. gate and review
are not authorization systems or security boundaries. Arbitype is an independent
open-source project and is not an official TypeSafe AI product.