luongnv89/claude-howto

This system is a comprehensive documentation and configuration repository for Claude Code, providing structured examples, skills, and plugins to extend its capabilities. It includes specialized tools for code review, DevOps automation, and static site generation, alongside localized documentation in Japanese, Ukrainian, and Vietnamese. The project also features scripts for measuring code complexity and quality, as well as utilities for managing permissions and tracking learning progress.

8 changes.

This period focused on expanding the project's documentation and tooling ecosystem, introducing comprehensive guides for advanced features, hooks, and plugins. It also added new automation capabilities, including a DevOps plugin, code refactoring skills, and a static website generator with corresponding test coverage.

9 changes.

This period focused on expanding internationalization support by adding comprehensive documentation and skill translations for Vietnamese, Ukrainian, and Japanese users. Concurrently, the project standardized its development environment by introducing structured Python build tooling and updated various plugins and skills to align with the requirements of Claude Code v2.1+.

The repository now includes a new 'refactor' skill under 03-skills, providing two Python scripts to help measure and improve code quality. The analyze-complexity.py script calculates metrics such as cyclomatic complexity, cognitive complexity, and maintainability index for Python, JavaScript, and TypeScript files. The detect-smells.py script identifies common code smells including long methods, duplicate code, large classes, and magic numbers, also supporting Python, JavaScript, and TypeScript. These tools are designed to help developers assess the impact of refactoring efforts.

python · high confidence

This change introduces the Vietnamese (Tiếng Việt) translation of the Hooks documentation and accompanying example scripts. The README in \vi/06-hooks\ now provides a complete guide to configuring and using Claude Code hooks—including command, HTTP, MCP tool, prompt, and agent types—along with a list of supported hook events. Additionally, example scripts such as \context-tracker.py\ (zero-dependency token usage tracking), \context-tracker-tiktoken.py\ (higher-accuracy tracking), \format-code.sh\ (auto-formatting on write), \pre-commit.sh\ (test gating), \security-scan.sh\ (secret detection), \validate-prompt.sh\ (dangerous operation blocking), \log-bash.sh\ (command logging), and \notify-team.sh\ (Slack/Discord/email notifications) are now available in Vietnamese, enabling users to implement these automation workflows with localized documentation.

vi/06-hooks · high confidence

The DevOps automation plugin now includes executable hooks and scripts to streamline Kubernetes deployments: pre-deploy and post-deploy Node.js hooks validate prerequisites (like kubectl availability and cluster connectivity) and wait for pod readiness, while bash scripts handle deployment, rollback, and health checks against API, database, and pod statuses. A new MCP configuration registers the Kubernetes server, and agent definitions (alert-analyzer, deployment-specialist, incident-commander) have been updated to use capitalized tool names (Read, Write, Bash, Grep) and the README now specifies a requirement for Claude Code 2.1+.

vi/07-plugins/devops-automation · high confidence

Added Japanese translations for the project's documentation, including slash command references, memory and skills guides, and issue/PR templates.

ja · high confidence

A new 'devops-automation' plugin (v1.0.0) has been added to the 07-plugins directory, providing slash commands for deploying to production or staging, rolling back deployments, checking system status, and handling incidents. The plugin includes specialized subagents (deployment-specialist, incident-commander, alert-analyzer) and integrates with Kubernetes via an MCP server configuration. It also brings over pre-existing automation scripts (deploy.sh, rollback.sh, health-check.sh) and hooks (pre-deploy.js, post-deploy.js) from the previous plugin location, now fully documented with support for Claude Code v2.1.220.

07-plugins/devops-automation · high confidence

Added a new Vietnamese (Tiếng Việt) translation of the advanced features guide, including the main README, planning mode examples, configuration examples, and a Python script for setting up auto-mode permissions. This provides Vietnamese-speaking users with localized instructions for planning mode, extended thinking, permission modes, and other advanced capabilities.

vi/09-advanced-features · high confidence

This update introduces comprehensive documentation for Claude Code's advanced capabilities, including planning mode, extended thinking, auto mode, background tasks, permission modes, and session management, along with configuration examples for various workflows like development, code review, and CI/CD. Additionally, a new Python script (\setup-auto-mode-permissions.py\) is provided to seed user settings with a conservative baseline of safe permissions for auto mode, allowing users to opt-in to additional permissions for editing, testing, git operations, and package management.

09-advanced-features · high confidence

The scripts directory now includes a static website generator (build\_website.py) that compiles markdown tutorials into a responsive, mobile-friendly site with dark mode, sidebar navigation, and self-hosted assets (Tailwind CSS, Mermaid, fonts). It also introduces an EPUB builder (build\_epub.py) for ebook generation, along with validation scripts for cross-references, external links, Mermaid diagrams, and markdown rendering correctness. A translation sync tool helps track outdated Vietnamese translations, and a vendor asset downloader ensures no external runtime dependencies for the generated site.

scripts · high confidence

This update adds a comprehensive set of example hook scripts and updated documentation for the 06-hooks module, aligned with Claude Code v2.1.126. The new examples include context usage trackers (with and without tiktoken), dependency vulnerability scanners, code formatters, bash command loggers, team notification senders, pre-commit test runners, pre-tool safety checks, security scanners for hardcoded secrets, session-end progress loggers, and prompt validators. The documentation has been rewritten to reflect the latest hook configuration options, including the new exec form for command hooks, HTTP hooks, prompt hooks, and the \if\ condition syntax for narrowing hook triggers.

06-hooks · high confidence

A new standalone HTML interface has been added to the local-progress directory to visualize learning progress. This page features a dark-themed dashboard with a global progress bar, a toolbar for importing data and resetting state, and an expandable list of modules. Each module displays its completion status and contains a grid of individual items, allowing users to track their advancement through the curriculum locally without external dependencies.

local-progress · high confidence

Added comprehensive Ukrainian translations for the Claude Code documentation and agent skills infrastructure. This includes translated READMEs for slash commands, memory management, and skills modules, along with localized content for specific agent skills such as blog drafting, brand voice, code optimization, and CI/CD setup. The update ensures that users interacting with Claude Code in Ukrainian have access to accurate guidance on slash commands, memory configuration, and custom skill creation.

uk · high confidence

The Code Review Specialist skill has been moved to a new location (03-skills/code-review-specialist) and updated to align with Claude Code v2.1.220. The skill now explicitly references and utilizes new supporting files: a structured review checklist (templates/review-checklist.md), a standardized finding template (templates/finding-template.md), and two Python scripts for quantitative analysis (scripts/analyze-metrics.py and scripts/compare-complexity.py). These scripts calculate code metrics like function/class counts and complexity scores, and compare complexity between code versions to aid in refactoring reviews. The skill description and trigger conditions have been broadened to include security analysis and performance optimization mentions. The underlying Python scripts have also been refactored to use double quotes and modern type hinting (dict instead of Dict).

03-skills/code-review-specialist · high confidence

The doc-generator skill has been moved to the 03-skills directory and updated to align with Claude Code v2.1.220. The skill metadata now includes specific compatible models (Claude Fable 5, Opus 5, Sonnet 5, etc.) and a last-updated date. The underlying Python script has been refactored to use modern Python typing (list\[dict\] instead of List\[Dict\]), removed unused imports (json, typing), and standardized string quotes to double quotes.

03-skills/doc-generator · high confidence

The PR Review plugin has been moved to the v07 directory and updated to require Claude Code 2.1+. The security-reviewer and test-checker agents now use the 'Bash' tool instead of 'diff' or lowercase tool names, ensuring compatibility with the new version's tooling standards. Additionally, the README now includes proper logo assets for dark mode support.

vi/07-plugins/pr-review · high confidence

The pre-review hook file has been moved from the 06-plugins directory to the 07-plugins directory to support a more optimal learning path for users navigating the plugin examples.

07-plugins/pr-review/hooks · low confidence

New pytest test suites have been added in the \scripts/tests\ directory to verify the correctness of the build tooling. The \test\_build\_epub.py\ file covers the EPUB builder module, including \BuildState\ management, \EPUBConfig\ defaults and overrides, input validation logic, and Mermaid diagram processing (sanitization and extraction). The \test\_build\_website.py\ file validates the static website generator, testing URL generation, link rewriting, directory exclusion rules, and heading-to-anchor conversion. Additionally, \test\_check\_cross\_references.py\ ensures the cross-reference checker correctly handles repo-root boundaries and missing files, while \test\_check\_markdown\_rendering.py\ verifies the markdown rendering validator against specific rules such as backtick usage, table pipes, and unmatched fences.

scripts/tests · high confidence

The repository now includes a structured Python environment for the \scripts/\ directory, featuring a \pyproject.toml\ that defines the core application (\claude-howto\) and its runtime dependencies (\ebooklib\, \markdown\, \beautifulsoup4\, \pillow\, \jinja2\). To ensure reproducibility, \requirements.txt\ pins these core packages to specific versions (e.g., \pillow==11.1.0\, \jinja2==3.1.4\). Additionally, \requirements-dev.txt\ and \pyproject.toml\ introduce development tooling, including testing frameworks (\pytest\, \pytest-asyncio\), code quality and linting tools (\ruff\, \mypy\, \bandit\, \pre-commit\), and their respective version constraints.

(dependencies) · high confidence

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

• First survey — no prior run to compare against. CAI 66.

• Code Health 63
• Architecture 100
• Maturity 81
• Readiness 60
• Security 80
• Accessibility 71

• 247 commits — 188 feature/other, 59 fixes

• (root) — 125 commits
• .claude/skills — 11 commits
• openspec/changes — 11 commits
• .github/workflows — 8 commits
• 01-slash-commands/README.md — 8 commits
• 03-skills/README.md — 6 commits
• (repo) — 5 commits
• 06-hooks/README.md — 5 commits
• 03-skills/blog-draft — 4 commits
• 02-memory/README.md — 3 commits
• 06-hooks/format-code.sh — 3 commits
• uk/03-skills — 3 commits
• uk/07-plugins — 3 commits
• .github/ISSUE_TEMPLATE — 2 commits
• 05-mcp/README.md — 2 commits
• 06-hooks/pre-tool-check.sh — 2 commits
• 08-checkpoints/README.md — 2 commits
• 09-advanced-features/README.md — 2 commits
• ja/07-plugins — 2 commits
• uk/01-slash-commands — 2 commits

• fix: Fixed a configuration error in the hooks Chinese guide
• fix: Revert "Merge pull request #149 from lilhawk7077/main"
• fix: ci: shift-left quality gates — add mypy to pre-commit, fix CI failures (#53)
• fix: fix(advanced-features): correct effort level model support (#96)
• fix: fix(ci): Add Sandboxing to cSpell dictionary
• fix: fix(ci): Add missing cSpell words and README sections for CI compliance
• fix: fix(ci): Add missing words to cSpell dictionary
• fix: fix(ci): Add sandboxed and pycache to cSpell dictionary
• fix: fix(ci): Use virtual environments instead of --system flag
• fix: fix(cli): correct agent priority order to match official docs (#98) (#100)
• fix: fix(docs): Fix broken links in resources and self-assessment skill
• fix: fix(docs): Remove dead marketplace link failing CI link checks
• fix: fix(docs): Update broken URLs, deprecations, and outdated references
• fix: fix(docs): Use tilde fences for nested code blocks in concepts guide
• fix: fix(docs): correct Claude Code version to 2.1.97 across all footers
• fix: fix(docs): escape inline-code backticks in !command`` references
• fix: fix(epub): embed SVG images instead of replacing with placeholders (#46)
• fix: fix(extended-thinking): correct fictional /think command to valid ultrathink keyword (#83)
• fix: fix(hooks): Windows Git Bash compatibility + stdin JSON protocol (#55)
• fix: fix(hooks): apply review fixes dropped by squash merge
• …and 227 more

• 0 containers · 1 bounded contexts · 0 dependency edges (baseline)

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

luongnv89/claude-howto was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.