User repository on GitHub
The skill-router mod: Jev picks and loads the skills each prompt needs.
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Integrations & Adapters
Skill Router — gecm0
A Claude Code mod that asks Jev one Noul question per skill for every prompt, ranks the skills above the threshold and passes them to the model as a suggestion. It never loads a skill itself, and a timeout, outage or missing key only skips the routing.
- Stars
- 3
- Language
- TypeScript
- License
- MIT
- Reading time
- 3 min
- Published by
- Community
- Author
- gecm0
- Last verified
- Oct 2026
Used inJev for Software & DevTools
- Languages
- TypeScript
- Platforms
- CLI
- Capabilities
- Open sourceRequires API key
Highlights
- One Noul question per skill, all in one request: 'is the user asking for this kind of task to be done now?', with a probability per skill.
- Over-threshold skills are ranked (default threshold 0.8, max 3) and shown above the prompt: 'Jev suggests · tdd 0.96 · animate 0.95 (240 ms)'.
- It only suggests; it never orders a skill loaded, because a mod cannot tell a skill that teaches from one that acts and every loaded skill costs context.
- Skips slash commands, prompts under 12 characters, and prompts not typed by you; already-loaded skills are not suggested again, and /clear forgets them.
- Fails open: a timeout, outage or missing key only skips routing; the default wait for Jev is 1500 ms.
Quickstart
bash
/plugin install jev --marketplace gecm0/jev-mod
/plugin install skill-router --marketplace gecm0/skill-router-modWatch out
README reviewed 2026-10-17; the mod needs the author's jev mod (gecm0/jev-mod) and a TypeSafe key, and the README gives no measured accuracy for the fit judgment, only the ~7,000-token, 200-450 ms example with about 90 skills.
Questions about this entry
Short answers drawn from the linked project or page.
A Claude Code mod that suggests the right skills for each prompt. On every prompt it asks Jev whether each skill fits, shows the ones over the threshold ranked above the prompt, and passes them to the model as a suggestion, never as an order.
Candidates are the skills the model can invoke itself, minus any the conversation already loaded, with user-only slash commands left out and duplicates offered once. Each skill gets one Noul question in a single request, so none fitting is a normal outcome.
The README explains that a mod cannot tell a skill that only teaches from one that delegates work or runs a subagent, and every loaded skill costs context. So Jev's picks arrive as a hidden note, and the model decides whether to load one.
It needs the jev mod and a TypeSafe key. Run /plugin install jev --marketplace gecm0/jev-mod, then /plugin install skill-router --marketplace gecm0/skill-router-mod.
It fails open: a timeout, an outage or a missing key only skips the routing for that prompt, and the band says why. The band clears when a prompt is not routed.
Every routed prompt sends the first 4,000 characters of your prompt plus each skill's name and one-line description to api.typesafe.ai, never a skill's full content. With about 90 skills that is roughly 7,000 input tokens and 200 to 450 ms per prompt.
In /plugin under Installed, skill-router, Configure options: the threshold (default 0.8), the maximum suggestions to show (default 3) and the timeout in milliseconds (default 1500).
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
Facebook: Intake Router keeps diagnosis out of the answer space at the front desk.
FacebookClinicnotes
X: OpenRouter Decision Model Rankings
X: herdr + jev, an agent router
X: A group chat bot with decision room routing
Other listings of this project
Forks and copies of this project in the directory. This entry is the canonical one.
Claude Code CLI that ranks every skill installed on your machine against a session goal using Jev, then tells the agent which one to load. Three possible answers: load this skill, nothing you have installed fits, or no skill is needed, and nothing else runs for the rest of the session.
Jev Agents & MCP#community#jevbest#catalog-import
Community
lomeshdutta/skill-routerPython
lomeshdutta/skill-routerPythonMore Jev resources like this
A Claude Code plugin that asks Jev, TypeSafe's System One model, which model tier each prompt needs — haiku, sonnet or opus — and switches automatically, and routes classification, scoring and yes/no judgments to Jev via scripts/jev.py.
Jev Integrations & Adapters#community#github#claude-code
Community
ricky94511/jevTypeScript
ricky94511/jevTypeScriptA Claude Code mod that sends each prompt to Jev through OpenRouter and turns the returned 0-3 score into a reasoning effort level, from low to xhigh. Shadow mode is on by default, so it only logs what it would set until you turn it off.
Jev Integrations & Adapters#community#github#claude-code
Community
gebeer/jev-effortTypeScript
gebeer/jev-effortTypeScriptA Claude Code plugin that picks Haiku, Sonnet or Opus and a reasoning effort for every prompt, step and subagent, with an effort cap, a prompt-cache guard and a status band; the repo is named jev-router and tagged typesafe.
Jev Integrations & Adapters#community#github#typescript
Community
dominicrico/jev-routerTypeScript
dominicrico/jev-routerTypeScriptA router that asks a classifier, Jev by default, how much reasoning each step of a coding session needs, then adds that effort to the model request without breaking the prompt cache. It ships as an OpenCode plugin, a Pi extension and a local proxy for Codex CLI.
Jev Integrations & Adapters#community#github#routing
A Claude Code plugin that picks the model tier and thinking effort for every turn with Jev and weighs each switch against what it costs, since moving a long conversation to another model rewrites its prompt cache and can cost more than it saves.
Jev Integrations & Adapters#community#catalog-import#logicrw
An unofficial TypeSafe Jev integration for Claude Code with hooks that suggest skills, classify task type, pick subagent model tiers and gate risky shell commands. It also ships jev-vision for yes/no, pick-one and bounding-box questions about images.
Jev Integrations & Adapters#community#github#javascript
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.