skill_optimizer_skill

Overview

This skill analyzes Claude Code skills for compliance, structure, and token efficiency. It automates detection of critical, recommended, and optional fixes and produces an optimized skill manifest that preserves intent while reducing token cost. The skill also produces a token budget comparison showing before/after savings.

How this skill works

It runs a static analysis script against a skill directory, then reads canonical checklist and pattern resources to cross-reference findings. The tool classifies issues into Critical, Recommended, and Optional, and creates concrete before/after fixes for each finding. Finally, it generates an optimized skill manifest constrained to allowed frontmatter fields and a concise body under the specified token budget.

When to use it

• Before submitting a skill for review to ensure compliance and efficiency
• When preparing a skill for production to reduce model invocation cost
• When consolidating multiple fixes into a prioritized remediation plan
• During audits to create a clear before/after token budget report

Best practices

• Run analysis on the canonical skill directory that contains the skill manifest and resources
• Address Critical issues first, then Recommended, and optionally Optional items
• Keep the manifest body imperative, directive, and under the token budget limit
• Move long reference material into resources to keep the manifest concise
• Use the token budget table to guide trimming of prompts and examples

Example use cases

• Automating a pre-release compliance check that enforces frontmatter field rules
• Reducing token usage for high-traffic skills to lower runtime costs
• Consolidating audit findings into an actionable remediation list with before/after examples
• Generating a compact, production-ready skill manifest that preserves original behavior