Token Audit

这是OpenClaw工作区的token消耗分析工具，核心功能为扫描工作区文件，估算GPT-4、Claude等6大模型的token数量与调用成本，识别冗余文件、臃肿上下文等成本驱动因素，并给出优化建议。典型场景是开发者需控制AI模型调用成本时使用，支持扫描整个工作区或指定路径，输出可选控制台文本或JSON格式，工具免费、零依赖，无需API密钥。

justincredible-tech

热度 26.71

最近更新 2026-05-20 03:28:24

分类：数据依赖：代码执行 / 读写文件应用场景：数据处理 / 财务 / 报告撰写

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add token-audit

Metadata

name: token-audit
version: 0.2.0
description: Workspace token consumption analyzer for OpenClaw agents. Scans workspace files, estimates token counts across 6 major models (GPT-4, Claude, Gemini, Llama, Mistral, Qwen), and surfaces hidden cost drivers. Identifies bloated context, redundant files, and optimization opportunities. Free, zero dependencies, no API keys.
author: Cael (@CaelMaximus)
license: MIT
tags: ["security","tokens","cost","optimization","workspace","caelguard"]

Token Audit

Analyze your OpenClaw workspace token consumption and costs across multiple models.

Usage

Scan your workspace

python3 scripts/token-audit.py

Scan a specific path

python3 scripts/token-audit.py --workspace /path/to/workspace

JSON output

python3 scripts/token-audit.py --json

What It Finds

Total token count per model (GPT-4, Claude 3, Gemini, Llama 3, Mistral, Qwen)
Per-file token breakdown sorted by size
Estimated costs per model at current pricing
Files that are disproportionately large vs their utility
Optimization recommendations

Part of Caelguard

Free, open-source agent security toolkit. See also: shellguard-scanner