skill-tracker

What This Skill Does

Skill Tracker is a powerful, privacy-first diagnostic tool designed for the OpenClaw ecosystem. It acts as a comprehensive usage analytics engine that monitors the performance and frequency of your installed agents and skills. By integrating directly into your Python and Node.js-based skills, it provides granular telemetry on call frequency, success/failure rates, and historical usage patterns. Because it operates entirely locally, Skill Tracker ensures that your operational data stays on your machine, preventing privacy leaks while providing the insights needed to optimize your AI workflow.

Installation

To install the Skill Tracker, open your terminal and run the command: clawhub install openclaw/skills/skills/aiyst1982/skill-tracker-aiyst. Once installed, the skill automatically registers its telemetry hooks. For custom skills, you simply import the tracker library and wrap your calls with the track() function. Python developers can utilize the provided module, while Node.js developers can leverage the provided commonJS/ES module to start logging performance metrics immediately without complex configuration.

Use Cases

Skill Tracker is essential for power users who want to move from manual management to data-driven agent optimization. Key use cases include:

1. Pruning Idle Skills: Identify which agents have not been used in weeks to reclaim system resources.
2. Reliability Auditing: Pinpoint specific skills that frequently fail during execution, allowing you to prioritize debugging efforts.
3. Usage Pattern Analysis: Determine which skills you rely on most heavily to streamline your workflow.
4. Trend Monitoring: Analyze if a particular skill's performance degrades over time due to external API updates or environmental changes.

Example Prompts

• "技能统计，查看我本周最常用的三个技能是什么？"
• "调用统计，分析一下百度搜索技能最近的成功率表现如何？"
• "查看技能使用情况，帮我找出最近调用失败次数最多的技能。"

Tips & Limitations

Tips:

• Regularly check the usage-log.jsonl file if you want to perform custom data processing outside of the dashboard.
• Use the config.yaml file to adjust the retentionDays setting to manage your storage footprint effectively.
• Integrate the tracker early in your development cycle to establish a baseline for your custom skills.

Limitations:

• The tracker introduces a negligible latency (less than 1ms per call), which is generally undetectable, but high-frequency loops (millions of calls) should be avoided.
• The tool currently does not provide complex visual graphing, focusing instead on text-based reports for faster command-line accessibility.