model-usage

What This Skill Does

The model-usage skill is a precision-engineered tool designed for power users who rely on the CodexBar CLI to track their AI expenditure and resource allocation. At its core, it parses complex cost JSON data generated by local CLI operations, distilling it into human-readable summaries. Whether you need a quick overview of your current model's usage or a granular breakdown across different AI providers like Codex or Claude, this skill provides the necessary visibility to manage your API budget effectively.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/ls18166407597-design/ag-model-usage Ensure that you have the latest version of the ClawHub CLI installed to prevent compatibility issues. Once installed, the skill will automatically detect your local auth-profiles.json to authorize requests.

Use Cases

This skill is indispensable for professional developers and data scientists who require strict cost monitoring. Use it during high-intensity coding sessions to ensure you are not exceeding daily limits, or when conducting batch processing tasks that involve multiple AI models. It is particularly helpful for teams attempting to optimize their resource allocation by identifying which specific tasks or models are consuming the most tokens.

Example Prompts

1. "Could you give me a summary of my current usage for the active model using CodexBar?"
2. "Show me a full breakdown of token costs across all models I've used this week."
3. "model-usage, check my total expenditure for Claude and compare it to my current Codex budget."

Tips & Limitations

To maximize the utility of this skill, ensure your CodexBar CLI is updated to the latest release, as changes in the JSON output format can affect parsing accuracy. Note that this skill is specifically architected for local CLI cost tracking; it does not directly query cloud billing dashboards, meaning it reflects local log data. Furthermore, ensure your environment variables are correctly set for auth-profiles.json access to avoid permission errors during the execution of diagnostic checks.