codebase-inspection

Inspect codebases with pygount to report LOC, language breakdown, and code-to-comment ratios.

1|Updated Jun 24, 2026

Search Tags:#codebase analysis #code metrics #pygount #language breakdown #lines of code #repository inspection

Ready to run

npx skills add https://github.com/maopujie10-sys/Bailongma --skill codebase-inspection-maopujie10-sys

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: codebase-inspection
Source: https://github.com/maopujie10-sys/Bailongma/tree/main/skills/hermes-skills/github/codebase-inspection
Command: npx skills add https://github.com/maopujie10-sys/Bailongma --skill codebase-inspection-maopujie10-sys

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pygount.

What problem does it solve? Answering questions about repository size, language composition, and code-to-comment ratios requires running line-counting tools correctly, and naive scans that crawl .git or node_modules can hang or produce misleading numbers. ## Core Features & Use Cases - Language Breakdown: Produces a summary table of files, code lines, and comment lines per programming language using pygount. - Safe Folder Exclusion: Provides ready-made --folders-to-skip lists for Python, JavaScript/TypeScript, and general projects to avoid scanning dependency directories. - Targeted Filtering: Supports --suffix filtering to count only specific languages and JSON output for programmatic consumption. - Use Case: A user asks how large a monorepo is and what percentage is Python versus TypeScript; the Skill runs pygount with proper exclusions and interprets the summary table, including pseudo-languages like generated and duplicate. ## Quick Start Use the codebase-inspection skill to count the lines of code and show the language breakdown of the repository at the current path.

Frequently Asked Questions about codebase-inspection

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

FAQPage Schema

How do I count lines of code in a repository?▼

Run pygount with --format=summary from the repository root to get a per-language table of files, code lines, and comment lines. Always pass --folders-to-skip to exclude .git, node_modules, and virtualenv directories.

How to count only specific languages with pygount?▼

Use the --suffix flag to filter by file extension, for example --suffix=py for Python only or --suffix=py,yaml,yml for multiple types. This narrows the scan and speeds up analysis on large repositories.

Why does pygount hang or take a long time on my repo?▼

Pygount crawls every directory by default, including .git, node_modules, and virtualenvs, which can contain hundreds of thousands of files. Always supply --folders-to-skip with dependency and build directories to avoid this.

Why does Markdown show zero code lines in pygount output?▼

Pygount classifies all Markdown content as comments rather than code, so Markdown files report zero code lines. This is expected behavior, not a bug in the scan.

What do __empty__, __binary__, and __duplicate__ mean in pygount results?▼

These are pseudo-languages pygount uses for non-code files: __empty__ for empty files, __binary__ for images and compiled artifacts, __generated__ for auto-generated files, and __duplicate__ for files with identical content.