typesafe

Hands the small judgments in a session to TypeSafe's Jev, a System One model that returns typed answers instead of…

좁은 판단(예/아니오·분류·0~N 점수)을 LLM 대신 jev 에게 넘겨 에이전트의 토큰과 시간을 아끼는 스킬. 타입이 고정된 확률과 캘리브레이션된 확신도를 건당 약 $0.00004, 300~500ms 에 돌려준다.

Room router backed by TypeSafe's Jev: when a room post addresses nobody, one System One request decides which agent…

Third-party skills enhanced with advisory TypeSafe Jev judgments: two-axis code review, prose humanizing (en/zh-TW),…

Asks TypeSafe's Jev which bulky tool results the session still needs, and replaces the rest with a bounded head plus a…

A curated collection of Jev skills, example projects, tools and integrations.

Screen prompts, skills and MCP tools with TypeSafe judgments and configurable reasoning models.

Jev-assisted, evidence-backed merge-queue decisions for rebase and full-test requirements.

Rule-based code review skills backed by the jev-flash-review MCP engine — review PRs, arbitrary code, or loop until…

High-level GOAL→JSON browser automation via TypeSafe Jev indexed controls. Prefer over screenshot computer-use for…

Suggests at most one skill for a request via Jev (TypeSafe cookbook pattern) before the agent loads anything.

TypeSafe Jev drives a real browser: one goal, an indexed action space, one systemOne decision per step. Exposes the…