Audit codebases for dead code, unused dependencies, and optimization opportunities.
npx skills add https://github.com/Capataina/agent-skills --skill code-health-audit
Please help me install this Agent Skill. Skill: code-health-audit Source: https://github.com/Capataina/agent-skills/tree/main/code-health-audit Command: npx skills add https://github.com/Capataina/agent-skills --skill code-health-audit
The code-health-audit analyzes the entire repository to identify opportunities to improve code health — dead code, unused dependencies, modularisation opportunities, hardcoded patterns extractable to algorithms, and opportunities for algorithm and memory-layout optimisations, as well as complexity hotspots, inconsistent patterns, API surface bloat, configuration drift, documentation rot, and test-coverage gaps, highlighting active risks. It reads project context and writes diagnostic tests (unit, integration, benchmarks, equivalence checks) to resolve uncertainty before findings are issued, and produces categorised plan files — never editing production source code.
Run the audit across the repository to generate a plan with findings and corresponding diagnostic tests.
High-intent search queries and answers about installing and using this skill.
A code health audit reads project context and conducts targeted research per substantive system to identify complexity hotspots, API surface bloat, and configuration drift, outputting categorized plan files with an obligation-evidence map.
Yes, the code health audit enforces non-functional-change discipline by never modifying production code, instead writing diagnostic tests like benchmarks and equivalence checks to resolve uncertainty before issuing findings in plan files.
Run the audit across your repository to generate categorized plan files in a dedicated plan folder, including an index, per-system findings, a live obligation-evidence map, and checkpoint documents for all identified optimisation opportunities.
The code health audit never modifies production source code, strictly producing plan files and diagnostic tests instead, meaning all identified code health improvements require separate manual implementation after the audit completes.
uoyoCsharpScan repository structure, detect languages and frameworks, and write a project context index.
ahravAnalyze and optimize Rust assembly output to fix compiler inefficiencies.
kriscendobotIncrease test coverage and identify unreachable code in a monorepo.
EmaRimoldiMap claims to experimental evidence and flag unsupported results in markdown.
