A monitoring utility that tracks real-time API quotas, remaining balances, and refresh schedules for high-performance AI models.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install ag-model-usage
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install ag-model-usage using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Model Usage?
The model-usage skill is a specialized utility designed for developers who need precise control over their AI consumption. By leveraging the CodexBar CLI and internal Google APIs, this skill provides real-time visibility into the usage metrics of models like Gemini and Claude. It is a vital part of the Openclaw Skills ecosystem, allowing users to avoid unexpected service interruptions by monitoring their remaining quotas and costs directly through their agent interface.
This skill acts as a bridge between the local development environment and cloud-based AI providers. It simulates official IDE client behaviors to fetch the most accurate account data, ensuring that users always have the latest information regarding their token limits and subscription boundaries.
Model Usage Use Cases
- Tracking daily or hourly quota consumption for Gemini and Claude models.
- Determining the exact local time when AI model limits will reset.
- Generating a detailed breakdown of model-level costs via CodexBar cost JSON files.
- Preventing workflow interruptions by checking remaining percentages before starting large-scale coding tasks.
How Model Usage Works
- The skill identifies the user's request for usage or cost data.
- It accesses the local auth-profiles.json file to retrieve the required OAuth tokens.
- A request is dispatched to the Google Antigravity (Cloud Code Assist) internal API, mimicking an official IDE client.
- The skill parses the returned JSON payload to extract specific quota metrics and refresh timestamps.
- The raw data is converted into a readable summary, with timestamps adjusted to the user's local timezone.
Model Usage Setup
To utilize this skill within the Openclaw Skills framework, ensure you have python3 installed and an active Google Cloud Code Assist account.
# Check for python3 dependency
python3 --version
# Ensure your authentication profiles are configured
# This skill reads from the standard auth-profiles.json location
Model Usage Data Schema & Taxonomy
The skill organizes usage data into a structured format for clear analysis:
| Attribute | Description |
|---|---|
| Model ID | The identifier for the specific AI model (e.g., Gemini-1.5-Pro) |
| Percentage | The remaining quota available expressed as a percentage |
| Refresh Window | The localized timestamp indicating when the next quota reset occurs |
| Auth Profile | The specific OAuth identity used for the API request |
Model Usage Advanced Features
- Direct integration with Google Antigravity for authoritative quota reporting.
- Automatic timezone conversion for all refresh estimates to ensure user accuracy.
- Support for multiple core models within a single summary view.
- Integration with local cost JSON files for historical usage analysis.
SKILL.md
Loading
Related Openclaw Skills
A-Share Stock Dossier
A professional-grade analysis skill that generates transparent, evidence-first dossiers for A-share stocks and portfolios using multi-source verification.
SymbolPicker
A specialized tool for implementing native-style SF Symbol pickers with color and style customization in SwiftUI applications.
SearchBar Skill
A comprehensive SwiftUI component skill for building highly customizable, cross-platform search bars with support for tokens and suggestions.
PagerKit SwiftUI Navigation Skill
An expert AI assistant for implementing PagerKit, a powerful SwiftUI library for highly customizable and cross-platform page-based navigation.
Netease Music Pusher
An automated music discovery tool that fetches and pushes personalized Netease Music daily recommendations and trending charts.
LM Studio Subagents
A specialized skill for offloading intensive or repetitive tasks from paid cloud providers to local LM Studio models.
Featured*
