Content
27%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is a comprehensive but excessively verbose technical debt analysis framework that reads more like a consulting playbook than an actionable Claude skill. It explains concepts Claude already knows (what code duplication is, what cyclomatic complexity means), uses hypothetical examples with made-up numbers rather than executable tool-based instructions, and packs everything into a single monolithic file. The content would benefit enormously from being trimmed to essential guidance and split across referenced files.
Suggestions
Reduce content by 70%+ by removing explanations of concepts Claude already knows (types of technical debt, what cyclomatic complexity is, etc.) and focus on the specific workflow steps and output format Claude should follow.
Add concrete tool usage instructions — specify how Claude should use file reading, grep, and other available tools to actually scan the codebase, rather than providing abstract category lists.
Split detailed templates (cost calculation templates, YAML metric dashboards, stakeholder report formats) into separate referenced files to improve progressive disclosure.
Add validation checkpoints: after the inventory step, Claude should verify findings with the user before proceeding to impact assessment and remediation planning.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose at ~300+ lines. Most content is generic technical debt knowledge that Claude already possesses — categories of debt, example cost calculations with made-up numbers, team allocation templates, stakeholder report templates. Almost nothing here is project-specific or adds knowledge Claude doesn't have. The skill reads like a textbook chapter, not a concise instruction set. | 1 / 3 |
Actionability | Contains concrete examples (code snippets, YAML configs, cost calculation templates) which provide some structure, but they are illustrative/hypothetical rather than executable against a real codebase. There are no actual tool invocations, CLI commands to run static analysis, or specific instructions for how Claude should use its available tools (file reading, grep, etc.) to perform the analysis. The guidance is more of a consulting framework than executable steps. | 2 / 3 |
Workflow Clarity | The 8 numbered sections provide a clear sequence, but there are no validation checkpoints or feedback loops. For a task involving codebase analysis and refactoring recommendations, there's no step to verify findings, validate metrics against actual code, or iterate on analysis. The workflow is a linear checklist without error recovery or verification gates. | 2 / 3 |
Progressive Disclosure | Monolithic wall of text with no references to external files and no bundle files. All content is inline in a single massive document. The detailed templates, examples, and category breakdowns could easily be split into referenced files (e.g., METRICS_TEMPLATES.md, REMEDIATION_PATTERNS.md). The sheer volume makes it hard to navigate. | 1 / 3 |
Total | 6 / 12 Passed |