What is this?
A Claude Skill that teaches Claude to measure a function's cyclomatic complexity and refactor the worst offenders instead of leaving tangled branch-heavy code behind.
A Claude Skill that teaches Claude to measure a function's cyclomatic complexity and refactor the worst offenders instead of leaving tangled branch-heavy code behind.Is this a standalone app, or does it plug into Claude?It is a Claude Skill — a packaged instruction set that Claude loads and follows, not a program you run yourself.
It is a Claude Skill — a packaged instruction set that Claude loads and follows, not a program you run yourself.How do I add this skill to Claude?For Claude.ai, download the .skill file from Releases (or zip the skills/cyclomatic-complexity folder) and upload it under Settings -> Capabilities -> Skills; it can also be added to Claude Code or uploaded via the Skills API.
For Claude.ai, download the .skill file from Releases (or zip the skills/cyclomatic-complexity folder) and upload it under Settings -> Capabilities -> Skills; it can also be added to Claude Code or uploaded via the Skills API.Does it need radon, eslint, or other analyzers installed first?It uses whichever complexity tool fits the codebase - radon, eslint, gocyclo, lizard, or a manual count - and respects your project's own linter thresholds if you have them configured.
It uses whichever complexity tool fits the codebase - radon, eslint, gocyclo, lizard, or a manual count - and respects your project's own linter thresholds if you have them configured.Does it just game the complexity score with tricks?No - it explicitly refuses to hide complexity in clever one-liners; complexity is moved into well-named, readable functions instead.
No - it explicitly refuses to hide complexity in clever one-liners; complexity is moved into well-named, readable functions instead.Who is this skill meant for?Developers cleaning up AI-generated code that works but has grown into deeply nested, hard-to-follow branching logic.
Developers cleaning up AI-generated code that works but has grown into deeply nested, hard-to-follow branching logic.When does the skill kick in during a session?It triggers automatically on refactoring, cleanup, and code review requests, or can be invoked directly by asking Claude to use it on a specific file.
It triggers automatically on refactoring, cleanup, and code review requests, or can be invoked directly by asking Claude to use it on a specific file.What do I get at the end of a refactor?Every refactor ends with a before/after complexity table so you can see exactly how much simpler each function became.
Every refactor ends with a before/after complexity table so you can see exactly how much simpler each function became.What refactoring techniques does it apply?It targets the worst complexity hotspots first using guard clauses, extract-function, lookup tables, and named predicates.
It targets the worst complexity hotspots first using guard clauses, extract-function, lookup tables, and named predicates.Which languages can it analyze?It works across languages by picking the right tool for the job - radon for Python, eslint for JavaScript, gocyclo for Go, or lizard as a multi-language fallback.
It works across languages by picking the right tool for the job - radon for Python, eslint for JavaScript, gocyclo for Go, or lizard as a multi-language fallback.What are similar projects?reclaim-code-entropy, simplify-codebase, linkedin-agent-skill, oil-ui, repowise, open-steps, skills, improve
reclaim-code-entropy, simplify-codebase, linkedin-agent-skill, oil-ui, repowise, open-steps, skills, improveWhich topics is it in?refactoring, claude-skill, code-quality, code-review, static-analysis, claude-skills, skill, skills, claude, markdown, agent-skills, ai-agent
refactoring, claude-skill, code-quality, code-review, static-analysis, claude-skills, skill, skills, claude, markdown, agent-skills, ai-agent