how to use claude code mods like a top 1% user, step by step: give this to your agent before everyone catches on👇 1. set up Jev connect Jev to the model registry you want to use. check that the connection works and the listed models are available. 2.... build your claude code mod open claude code 2.1.287 or later and paste this prompt: “build a mod called run-ledger. load plugin-authoring and use the API types for my installed version. create a dashboard that shows: - estimated cost per run, model, and source plugin where known - which model handles each task - each subagent’s status, latest action, and elapsed time include token counts, cache usage, and reported retries. count each request once. keep background tasks linked to their original run. use dated prices. label costs as API estimates, not subscription charges. show unknown when data is missing. connect the mod to my existing Jev setup. give Jev the task, available models, prices, budget, and relevant past results. ask it to recommend a model and explain why. start with recommendations. make automatic routing optional for eligible subagents. show the recommended model, actual model, result, and cost. include Jev’s own cost. keep state across hot reloads. add details and export. the dashboard itself must make no model calls. validate the plugin. test rendering, costs, attribution, and duplicate counting. give me steps for a live test.” 3. test the setup allow hot reload when prompted. run a simple task, a subagent task, and a mod-triggered model call. check that the dashboard updates, each request counts once, and Jev’s recommended model can actually run the task. 4. install the working mod ask claude to copy it out of the temporary folder and install it as a persistent plugin. see the cost. choose the model. check the result.show more

@Av1dlive make sure to version control your claude code. saves headaches later on!

thanks for sharing this setup

You're welcome, brother. What did you end up actually liking about this?

this is very useful my man

Why route when you can just use latest models lol

A live test with deliberately missing price data would catch more than a happy-path run.

Claude has improved significantly when it comes to working with agents

this needs to be mastered by every claude code builder

Tnx for this

can the dashboard export all run-ledger data as csv?

Yes, it actually can. I'm going to open-source it real soon. The video you just saw is my own, a visual rendition of how it looks on my Harness, which I have built. I'm going to create a version that will work on Claude Code.

recommendations first and automatic routing after the right order for any system that handles money

It's always like that, even for me. Jev is a big part of decision-making because we need quick decisions to be made when we are working with AI, given that we are at the frontier.

at the frontier you have to decide fast good that at least one model does it without panic

Thanks for sharing this!

I just gave this to my agent.. thanks

@DeathHomelander perf!

wow that routing panel looks clean and this whole guide is really well made

Thanks, Argona. I appreciate you, bro. I think Jev is a real change maker, and I think nobody has paired Jev with Claude Code mods till now. Claude Code mods allow you to decide whether this mod should even exist, and Jev can play that role depending on the need. If you define the needs well enough, it can create it.

alpha knowledge

this kind of cost tracking could make agent workflows much easier to manage. bookmarking this to experiment with model routing.

Showing model roater, agent graph is a good one

what's the biggest challenge you've faced when setting up Jev with the model registry?

For step 1, I’d use GLiDE as the router. It beats Jev by 6.9 points on Decision Index and returns probabilities for the available choices. The run ledger would also let you compare routing mistakes rather than just the bill.

Tracking subagent costs and status in real time is essential