model-usage

Aggregate CodexBar cost JSON into per-model cost summaries.

Updated Feb 8, 2026
One-click install
npx skills add https://github.com/nomad3/openclaw-k8s --skill model-usage-nomad3
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: model-usage
Source: https://github.com/nomad3/openclaw-k8s/tree/main/package/skills/model-usage
Command: npx skills add https://github.com/nomad3/openclaw-k8s --skill model-usage-nomad3

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill eliminates the manual effort of turning CodexBar local cost logs into a clear, per-model view of usage and cost.

Core Features & Use Cases

  • Per-model cost summaries: Produces a cost-only breakdown for either the current (most recent) model or an all-model list from CodexBar cost JSON.
  • Current-model selection logic: Chooses the highest-cost model within the latest daily entry that includes modelBreakdowns, with a fallback to the last item in modelsUsed.
  • Scriptable automation: Accepts live data from codexbar cost --provider <codex|claude> --format json or an input JSON file/stdin, and can output text or JSON.
  • Use Case: Use it to quickly answer “What model did I spend the most on recently?” or “How did my Codex/Claude model costs distribute across models over time?” for local cost-only reporting.

Quick Start

Run this Skill to get an all-model per-model cost breakdown from your local CodexBar cost logs by telling the assistant: “Summarize CodexBar local cost usage for all Codex models and return the per-model totals.”

Frequently Asked Questions about model-usage

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I summarize CodexBar cost logs by model?▼

You can summarize CodexBar cost logs by model by feeding the JSON output from `codexbar cost --format json` into this tool, which aggregates `daily[].modelBreakdowns` to produce a per-model cost breakdown.

What is the best way to see which model incurred the highest cost recently?▼

To see which model incurred the highest cost recently, the tool identifies the current model by selecting the highest-cost entry in the latest daily row's `modelBreakdowns`, with a fallback to the last item in `modelsUsed`.

Can I parse CodexBar JSON cost data from a local file instead of the CLI?▼

Yes, you can parse CodexBar JSON cost data from a local file or standard input, allowing you to process saved cost logs and aggregate per-model usage without running the live CLI command.

Does this tool work with both Codex and Claude provider cost JSON?▼

Yes, it works with both Codex and Claude provider cost JSON, accepting data from `codexbar cost --provider <codex|claude> --format json` to aggregate local cost-only reporting across models.

How do I get a full per-model cost breakdown across all historical data?▼

To get a full per-model cost breakdown across all historical data, instruct the tool to aggregate all daily entries from the CodexBar cost JSON, outputting a comprehensive list of per-model totals in text or JSON format.

Why does my current model selection show a different model than expected?▼

Current model selection might differ from expectations because the logic prioritizes the highest-cost model within the latest daily entry containing `modelBreakdowns`, defaulting to the last item in `modelsUsed` only as a fallback.