skill-audit

Analyze usage logs to classify skills and flag duplicates for audit reports.

Updated Apr 24, 2026
One-click install
npx skills add https://github.com/Skobyn/g.c.l.a.w --skill skill-audit-skobyn
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-audit
Source: https://github.com/Skobyn/g.c.l.a.w/tree/main/skills/skill-audit
Command: npx skills add https://github.com/Skobyn/g.c.l.a.w --skill skill-audit-skobyn

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Analyze GClaw skill and capability usage to identify what's working, what's dead weight, and what should be pruned or improved. Run after collecting a few weeks of usage data. Also audits agent capability usage — which tools, crons, and specialist agents are actually pulling weight versus dead weight.

Core Features & Use Cases

  • Collects and analyzes usage events (from usage_events Firestore) to surface actionable insights about skill and capability effectiveness.
  • Produces a written audit showing frequency, recency, agent distribution, and trends; flags duplicates and underperformers; cross-references the skill registry with actual usage.
  • Helps product teams and agents decide which skills to REMOVE, IMPROVE, KEEP, or PROMOTE based on data.

Quick Start

Use this skill to generate an audit after accumulating several weeks of usage data, then review and apply the recommendations in the admin UI.

Frequently Asked Questions about skill-audit

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

FAQPage Schema
How do I audit skill usage logs to identify underperforming agents?

To audit skill usage, you analyze usage events from Firestore to classify skills as active, occasional, dormant, or never-used, and flag duplicates or underperformers across agents. The audit report gets saved to audits/skills for review.

When should I run a skill registry audit to prune dead weight?

You should run a skill registry audit after accumulating several weeks of usage data. This timeframe ensures enough events exist in Firestore to accurately classify which skills to keep, prune, consolidate, or promote based on actual recency and frequency.

What is the best way to classify dormant skills versus active ones?

The best way to classify dormant versus active skills is cross-referencing the skill registry with actual usage dashboards and Firestore events. This classifies skills by frequency, recency, agent distribution, and trends to identify dead weight.

How do I flag duplicate skills across multiple agents?

You flag duplicate skills by cross-referencing the skill registry with actual usage data from Firestore usage_events. The audit process identifies overlapping capabilities across agents and includes them in a written report for consolidation decisions.

Can I analyze agent capability usage for crons and specialist tools?

Yes, you can analyze agent capability usage for crons and specialist tools. The audit process evaluates which tools are pulling weight versus dead weight by reading usage_events in Firestore and cross-referencing with the skill registry.

Does skill auditing require a minimum amount of Firestore usage data?

Skill auditing requires several weeks of accumulated usage data in the Firestore usage_events collection. Without sufficient historical events, the audit cannot accurately determine recency, frequency trends, or classify dormant and never-used skills.