openclaw-jev-trigger
OpenClaw plugin and CLI that turn a plain-language condition into a scheduled trigger, waking the model only when Jev says it is true.
Overview
openclaw-jev-trigger lets you write OpenClaw automation triggers in plain language. OpenClaw's gateway-owned scheduler runs a small script on every tick that returns fire false or fire true with a message. This plugin adds a jev_when tool and a CLI that writes the trigger script from --when and --not-when conditions, such as CI turned red but not while CI is still running.
Each tick asks the agent's decisionModel (hosted Jev by default, or a local Kev or CPU ONNX classifier) one Noul question about what the watcher observed; only a yes at 0.7 or above wakes the conversation model. On 76 synthetic watcher ticks the README reports Jev right on 75 with no false wake-ups, 231 ms median and about $0.000016 per check, roughly $0.14 a month for a 5-minute watcher, against 87% for first-try JavaScript rules. The README states the code and benchmark were written by an AI coding agent and checked on a real OpenClaw gateway.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Is the --when condition true (and the --not-when condition false) for this observation? | yes · no |
How it works
- Generate a trigger script with --when, --not-when and --exec.
- Schedule it as an OpenClaw automation.
- Each tick runs the watcher and asks the decision model.
- Fire only when the probability is at least 0.7.
Strengths
- Quiet ticks cost no model tokens.
- Handles conditions awkward in JavaScript.
- Swappable decision backend.
Caveats
- Benchmark data is synthetic.
- Requires OpenClaw with a configured decisionModel.
Getting started
openclaw-jev-trigger --when "CI turned red" --not-when "CI is still running" --exec "gh run list --limit 1 --json status,conclusion" > ci-red.js, then openclaw automations add --every 5m --trigger-script ci-red.js.At a glance
- Type
- Plugin
- Status
- beta
- Language
- TypeScript
- Runtime
- OpenClaw plugin, CLI
- Has code
- yes
- Has tests
- yes
Use cases and integrations
How curated lists describe it
Agent automation: OpenClaw plugin and CLI that turn a plain-language `--when` / `--not-when` condition into a scheduled trigger script, asking Jev one `Noul` per tick through OpenClaw's `decisionModel` and waking the conversation model only when the condition becomes true at 0.7 or above; on 76 synthetic watcher ticks Jev was right on 75 with 0 false wake-ups at 231 ms p50 and about $0.000016 per check, against 87% for first-try JavaScript rules.
yibie/awesome-jev
Frequently asked questions
What does Jev decide each tick?
Whether the plain-language condition is now true for what the watcher saw.
How much does it cost?
About $0.000016 per check, roughly $0.14 a month for a 5-minute watcher, per the README.
Can I use a local model?
Yes, a local Kev or CPU ONNX classifier via the decisionModel setting.
Related projects
AutoGPT agent platform, listed as adding Jev decision blocks for typed routing, filtering, scoring and confidence-gated next actions.
ProjectproductionPythonMulti-agent orchestration layer for Claude Code with opt-in Jev hooks for fast judgment points, decision caching and egress controls.
PluginproductionTypeScriptRuns Claude Code on Opus 5.5 with Jev re-picking reasoning effort after every tool batch, via per-message effort so the cache survives.
ProxybetaTypeScriptEvent-driven multi-agent network where Jev watches live state and assigns pending tasks to idle general-purpose agents.
ProjectprototypePythonPulumi example where Jev picks the best cloud region and modules per user location while code does pricing math and provisioning.
ProjectprototypeTypeScriptDeepSeek Harness plugin for budgeted, read-only GPT 'teacher' consults; Jev may only suggest when a consult is worthwhile.
ProjectprototypeJavaScript