A model router that chooses a model size for each message and can be switched on, switched off, or checked for status. It sends messages to Jev, a service that estimates the smallest suitable model.

Use it to route messages by estimated difficulty, inspect routing decisions, and manage the router for Claude Code, Codex, or opencode.

It removes the need to choose a model size manually for every message. It can also show how messages were classified and what they cost when that information is available.

jev — Skill for Claude Code by yorrick/agent-skills

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add yorrick/agent-skills --skill jev

Clone the repo

git clone --depth 1 https://github.com/yorrick/agent-skills

Made for: Claude Code.

Or install jev-router, the plugin that ships this one along with the rest of its 1 skill, 4 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/yorrick/agent-skills/jev/github.svg)](https://agentmods.dev/skills/yorrick/agent-skills/jev)

Your own site

<a href="https://agentmods.dev/skills/yorrick/agent-skills/jev"><img src="https://agentmods.dev/badge/skills/yorrick/agent-skills/jev/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/yorrick/agent-skills/jev"><img src="https://agentmods.dev/badge/skills/yorrick/agent-skills/jev.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 189 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,305 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00189	$0.01305
Opus 5.5	$0.00076	$0.00522
Sonnet 5.5	$0.00038	$0.00261
Haiku 4.5	$0.00019	$0.00130

Measured 19d ago against content hash 3ab20b7c062e, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

jev scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 19d ago.

The scan reads SKILL.md. This mod also ships 11 executable files (scripts/delegation.py, scripts/fork_check.py, scripts/jev_router.py, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

jev-router/skills/jev/SKILL.md · 36 lines

How it starts

The opening of the file, as written. The whole thing — 36 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Jev router switch

Run the router script in this skill's scripts/ directory with the user's word: on, off or status (no word means status).

uv run --quiet --script <this skill's directory>/scripts/jev_router.py status

Show the user the script's output as it is. It already carries the privacy reminder when the router is on. Do nothing else.

on needs, once, the shell file that exports OPENROUTER_API_KEY: add --key-file <path>. The path is remembered for later ons. If on fails because no key file is known, ask the user for the path. Never search for keys and never print one.
classify "<message>" ... sizes messages without routing anything and prints a table of what Jev picked and how sure it was. Add --harness codex or --harness opencode to use those harnesses' sizes.
JEV_ROUTER=off in a run's environment turns the router off for that run, whatever the switch says; JEV_ROUTER=on opts a headless run (claude -p, codex exec, opencode run) in, such as an eval.
mode shadow|capture|live (runs jev_router.py mode ...) sets the Claude Code cache-aware router's mode: shadow decides and logs but tells the session nothing, capture also saves a snapshot of each job it would delegate, and live tells the session to hand the job to a fresh subagent. Add --dir <path> with capture to change where snapshots are saved; a path with a space or a non-ASCII character is refused.

Fork check

The fork check replays the jobs the router selected in capture mode, both kept and delegated, to measure real cost, time and quality before turning on live mode. Run its commands from skills/jev/scripts/fork_check.py, in this order: check (restore-check new snapshots), list (snapshots and what each real turn did outside the machine), mark (mark a snapshot safe or skip to replay, or inconclusive to take a stuck job out of the report), replay (restore a marked snapshot into two clones and run keep and delegate; it also resumes a job an interrupted replay left behind), judge (a blind Codex verdict on the two results, whose copies leave out ignored files outside dependency folders and virtualenvs, and any copy of a restored secret such as .env; a result holding a link out of the repository is not judged), publish (push both results to a private copy and open the comparison PR), report (the fork check's pass or fail over its 20 jobs), and shadow (shadow-mode decisions against what really happened).

Read the full file on GitHub · 36 lines

Files

What ships with it

14 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

19d ago Changed · +8 lines · +81 tokens per session 3ab20b7c062e
21d ago Changed 63ca9f5ae001
22d ago First seen · 28 lines · 108 tokens per session scan A 384eaf1c2a82

Subscribe to this mod's changes

jev is a skill published in the GitHub repository yorrick/agent-skills (11 stars, last pushed 11d ago), licensed MIT. It adds 189 tokens to every session and 1,305 once invoked, about $0.0008 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-28.

Data and AI

agent-platform-model-registry

Agent Platform Model Registry Management. Use when you need to upload, list, describe, update, or delete machine learning models (and their versions) in the Agent Platform Model Registry. Don't use for model training, model deployment to endpoints, or managing non-Agent Platform models.

google/skills · 60 tokens

foundry-config-setup

Resolve missing setup caused by a hardcoded Foundry project endpoint or model in a sample. Use when a sample fails because it uses a placeholder/hardcoded projectendpoint (for example "https://your-project.services.ai.azure.com") or a hardcoded model instead of reading them from the environment.

microsoft/agent-framework · 65 tokens

google-cloud-solution-agentic-analytics-spark-knowledge-catalog

Discovers requirements and designs an end-to-end governed agentic analytics solution using Knowledge Catalog and Managed Service for Apache Spark (Lightning Engine). Use when designing data science and analytics workflows across structured and unstructured distributed data (including in S3, Azure Blob, AlloyDB, and…

google/skills · 132 tokens

training-check

Interactively monitor training metrics from the current Codex session, periodically checking WandB or fallback logs for NaN, divergence, plateaus, and broken runs.

wanshuiyin/Auto-claude-code-research-in-sleep · 35 tokens

nemo-automodel-launcher-config

Configure NeMo AutoModel job launches for interactive runs, Slurm clusters, and SkyPilot cloud execution.

NVIDIA/skills · 30 tokens

finetuning

Fine-tune models on Microsoft Foundry using SFT (supervised), DPO (preference), or RFT (reinforcement with graders). Covers dataset preparation, training job submission, deployment, and evaluation. USE FOR: fine-tune, SFT, DPO, RFT, training data, grader, distillation, fine-tuned model, training job, large file…

microsoft/skills · 132 tokens