A community Claude Code Templates mod that uses Jev to suggest subagent models and reasoning levels.
Evaluates task tier, reasoning needs and production risk; local policy maps results to invocation settings.
Makes routing rules and per-request choices explicit and configurable.
A davila7 community project, not an official Anthropic repository or mod. Aggregate savings and safety effects were not verified here.
Route a coding task to a configured model tier and reasoning effort.
Material analysed：
jev-spike.mjs explicitly contrasts native noul/confidence with Gateway boolean/probabilities. Its confidenceOf helper derives a maximum probability only when a native confidence field is absent.
Inspect the supporting source ↗The module combines tier and reasoning-demand results with configured mappings. The example model identifiers and thresholds are a routing policy, not proof that a tier can solve a given repository task.
Inspect the supporting source ↗Use it to make routing choices inspectable and configurable. Measure final task success and total cost, including misroutes, before accepting savings claims.
Compare the two provider adapters and their confidence/risk readers, then inspect jev-model-router.ts for how each outcome selects model and effort.