jevbooks

← All projects

wakegate

shitianfang/wakegate

Ask Jev if a sleeping agent's wakeup is worth an LLM turn.

Agent Decisions54%Runner-up: Classification & RoutingEnter / exit thresholds

What it is

A fail-open wake gate for long-running agents on Workers, Durable Objects and Node. Before a sleeping agent's LLM is resumed on a timer or incoming event, wakegate asks Jev whether the wakeup is worth a full LLM turn.

How it uses Jev

Jev answers a three-way Choice (wake, not yet, unrelated) against the agent's own sleep note, event and optional observation. Code skips the wakeup only when Jev puts less than 0.2 on wake; every other path wakes the agent.

Primitives:choice

Technique worth stealing

Put one Jev call in front of the LLM resume and fail open on every uncertain path.

Try it

npm install github:shitianfang/wakegate ai; run eval with AI_GATEWAY_API_KEY=... npm run eval

View on GitHub

judged by Jevjev-1.13.0

Evidence

Each line is one question put to Jev about the README. ≥ 0.60 reads as yes, ≤ 0.40 as no; in between Jev is not making a call.

  • Jev-centricyes0.95
  • Shows a System One patternyes0.69
  • Handles uncertaintyyes0.90
  • Measuredyes0.97
  • Runnableyes0.85
  • Worth recommendingunclear0.56
  • Model replicano0.03
  • Problem scopescore on a 0–2 scale1.20
  • About Jevyes0.98

Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 20 Sept 2026.