Jev System Architect Skill

September 17, 2026 · View on GitHub

A system-architecture skill for TypeSafe AI's Jev / System One. Continuously asks: Where does this system contain fuzzy semantic judgment that should become a small Jev decision primitive?

TypeSafe already publishes an official skill that gives coding agents current API/SDK context and implementation guidance. The missing layer is architecture: keep deterministic logic, control flow, and side effects in code; use Jev for narrow structured judgments; batch independent questions; and route uncertainty explicitly.

Install both: TypeSafe's official skill and this one.

Install

Copy into your agent skills folder (Cursor / Claude Code / compatible harness):

# from this repo
cp -R skills/jev-system-architect ~/.cursor/skills/jev-system-architect
# or wherever your agent loads skills from

Or clone and point your agent at skills/jev-system-architect/SKILL.md.

What this skill is for

Use when a system contains unstructured data, brittle parsing, prompt-to-JSON workflows, semantic routing, classification, ranking, verification, guardrails, extraction from known candidates, or other places where deterministic code needs bounded human-like judgment.

It helps you determine:

  1. Whether Jev belongs in a system at all
  2. Exactly where the Jev boundary should be
  3. Which judgments should become Choice, Score, or Noul questions
  4. What state each judgment should see
  5. Which questions should run together
  6. How ordinary code should compose the results
  7. How uncertainty should change system behavior
  8. How the integration should be evaluated before production use

This is primarily an architecture and decomposition skill.

For exact SDK syntax, request/response fields, models, limits, or version-specific API behavior, read the current TypeSafe documentation and use TypeSafe's official agent skill when available. Do not invent API fields.

Core mental model

APPLICATION STATE
      ↓
bounded semantic questions
      ↓
     JEV
      ↓
typed judgments + probabilities
      ↓
DETERMINISTIC APPLICATION CODE
      ↓
route / rank / verify / ask / escalate / execute

The model supplies programmable judgment. Code owns control flow, permissions, calculations, persistence, transactions, side effects, tool execution, authorization, and policy enforcement.

  • TypeSafe AI / Jev docs — source of truth for API and SDK details
  • TypeSafe official agent skill — implementation guidance

License

MIT — see LICENSE.

Built for public reuse by Ripe Avocado.