I kept watching coding agents burn context on decisions that aren't hard - triage 400 tickets, tag 600 files, route to one of six teams. jev-mode moves those verdicts to a typed-judgment model. I A/B'd it: 78% fewer tokens, 16x less work-attributable input, accuracy 96.1% vs 93.7%. Python, no deps, MIT.
The link points at the commit we read, so the line number still holds.
This one cannot run here
Its question set is assembled at runtime, or never written out literally in the code, so there is nothing to lift. The source link above will show you.
Other projects in this category
JevRouter ▶ — A lightweight Jev-powered router for models, tools, and subagents
pi-jev ▶ — TypeSafe Jev as a decision layer for the Pi coding agent: a measured tool-call gate plus jev_ask for typed, calibrated answers
stanley-code ▶ — Bounded TypeSafe Jev workflows for coding agents.
jev-sift ▶ — Classify first. Read selectively. A portable agent plugin and MCP tool for batch text classification.
grok-bot-jev ▶ — Connect TypeSafe Jev to Grok Bot as a cheap decision layer - usage gates, skill template, examples
Jevbridge ▶ — ACP and MCP adapter that bridges TypeSafe Jev with any LLM — computer use and typed decisions alongside Codex, Claude, Grok, and OpenCode.
jev-mcp ▶ — MCP server that puts TypeSafe Jev on the coding loop in Cursor, Codex, and any MCP client
pi-jev ▶ — Semantic tool routing and typed System One decisions for the Pi coding agent using TypeSafe Jev