Claude Code Best Practices
From vibe coding to agentic engineering: a structured guide to configuring and mastering Claude Code.
- Source repo
- shanraisshan/claude-code-best-practice
- Stars
- ★ 68k
- Last updated
- today
- License
- MIT
- Primary language
- HTML
- FA score
- 25/100 · Major gaps
At a glance
- How it runs
- Works with
- Platform-specificClaude Code
- Cost
- Free software; you pay for model usage
- Setup effort
- Low · running in minutes
- You'll need
- Typical use
- Developers wanting to systematically configure Claude Code (subagents, commands, skills, hooks, MCP) can use the best-practice notes and implementation examples.
- Not a fit if
- Teams wanting one-click setup or automated installation
- Developers not using the Claude Code ecosystem
- Teams needing content stable across Claude Code versions
- Source review
- 25/100 · Major gaps 8 safety controls not found
What does this agent do, and when should you use it?
This repository is a comprehensive best-practices guide for Claude Code, maintained by shanraisshan, aimed at helping developers transition from casual vibe coding to rigorous agentic engineering. It covers core Claude Code concepts such as subagents, commands, skills, hooks, MCP servers, settings, memory, checkpointing, and CLI startup flags, each with best-practice notes and implementation examples. It also includes comparative analyses of various development workflows (e.g., Superpowers, Everything Claude Code, etc.), and distills 83 practical tips from Claude Code creator Boris Cherny and community experts covering prompting, planning, context, session management, CLAUDE.md, agents, commands, skills, hooks, workflows, git/PR, and debugging. Additionally, it documents cross-model collaboration approaches (via plugins, MCP, or routers) and curates relevant videos and podcasts. The project is MIT-licensed, actively updated, and serves as a reference manual for developers aiming to deeply master Claude Code.
The repository provides structured documentation and examples demonstrating how to configure Claude Code components: subagent definitions (.claude/agents/*.md), commands (.claude/commands/*.md), skills (.claude/skills/*/SKILL.md), hooks (.claude/hooks/), MCP settings (.mcp.json), settings files (.claude/settings.json), and memory files (e.g., CLAUDE.md and .claude/rules/). It includes an orchestration workflow example (weather-orchestrator) that demonstrates the command→agent→skill pattern. Moreover, it references numerous external workflows (e.g., Superpowers, Everything Claude Code) and provides copyable commands like claude and /weather-orchestrator. It curates best-practice tips from the Claude Code team (Boris, Thariq) and the community, and documents cross-model integration options (e.g., via router or MCP to call other models).
- Developers wanting to systematically configure Claude Code (subagents, commands, skills, hooks, MCP) can use the best-practice notes and implementation examples.
- Teams designing complex development pipelines (like research→plan→execute→review→ship) can draw from the workflow comparisons and orchestration examples.
- Developers facing context management issues (e.g., context rot, long sessions) can apply tips on /compact, /clear, /rewind, and subagent usage.
- Developers wanting to integrate Claude Code with other AI models (e.g., Codex, Gemini) can refer to the cross-model workflow section.
- Developers writing high-quality CLAUDE.md and skill files can follow the guidance and recommendations in the repo.
- Developers tracking the latest Claude Code features and community trends can explore the 'Hot' section and subscription list.
How do you install or deploy this agent?
This repository is a collection of documentation and examples; no installation is required. Users must first install the Claude Code CLI (e.g., via npm install -g @anthropic-ai/claude-code or per official docs), then clone the repo: git clone https://github.com/shanraisshan/claude-code-best-practice.git. The templates and examples (like the .claude/ directory structure) can be copied into your own project.
How do you use this agent?
After cloning, browse the docs. To use the built-in orchestration example, run /weather-orchestrator in Claude Code to experience the command→agent→skill flow. Reference the 'best-practice' and 'implementation' directories to copy subagent, command, skill, etc. files into your project. Read the 'Tips and Tricks' section and apply applicable tips to your daily workflow.
What are this agent's strengths and limitations?
- Comprehensive and well-structured, covering core Claude Code concepts and advanced features with best-practice notes and implementation examples.
- Aggregates 83 practical tips from Claude Code creator Boris Cherny and community experts, offering high practical value.
- Provides comparative analysis of multiple well-known development workflows (e.g., Superpowers, Everything Claude Code) to help users choose a methodology.
- Highly dependent on the Claude Code ecosystem; some content (e.g., specific commands or features) may become outdated as versions evolve, requiring continuous attention.
- Primarily serves as a knowledge base; users must understand and apply it themselves, with no one-click configuration or automated installation tools.
- Some content (e.g., tips and videos) links to external resources that may break, and is not fully backed up within the repository.
How does this agent compare with similar options?
The repository, in its development workflows section, compares several well-known Claude Code workflow repositories, including Superpowers (obra/superpowers), Everything Claude Code (affaan-m/everything-claude-code), Matt Pocock Skills (mattpocock/skills), gstack (garrytan/gstack), Spec Kit (github/spec-kit), agent-skills (addyosmani/agent-skills), Get Shit Done (gsd-build/get-shit-done), OpenSpec (Fission-AI/OpenSpec), BMAD-METHOD (bmad-code-org/BMAD-METHOD), oh-my-claudecode (Yeachan-Heo/oh-my-claudecode), etc., providing step-by-step flows and star ratings to help users choose based on their needs.
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| Claude Code Best Practices This agent | 25 · Major gaps | CLIFree + model costs | ★ 68k | today | HTML | Claude Code |
| Claude Code: Everything You Need to Know | 50 · Major gaps | CLIFree + model costs | ★ 3.4k | 1d ago | JavaScript | Claude Code |
| MiMoCode | 50 · Major gaps | CLIFreemium | ★ 14k | today | TypeScript | ChatGPT · Codex · Claude Code |
| Whale | 51 · Major gaps | CLIFree + model costs | ★ 934 | 5d ago | Go | — |
How does FollowAgents rate this agent?
Why each dimension lost points
The evidence shows the repository is primarily documentation and configuration, with no actual code or executable files, so there is no basis to assess least privilege, user confirmation, data flow transparency, sensitive data handling, dependency security, external effects, rollback, or source attribution. All trust criteria score 0 because there is no evidence to support them.
Self-consistency: The features listed in the README are largely consistent with the repository structure, but there are broken links (e.g., '!/tags/...') indicating incomplete or erroneous documentation. Dependency availability: No dependencies are listed, so cannot be assessed. Failure messages: No documentation of error handling or failure messages.
Audience and scenarios: The README clearly targets Claude Code users, providing various workflows and best practices, suitable for developers. Capability boundaries: The documentation describes various features but does not clearly state limitations or boundaries. Trigger precision: Provides usage instructions for commands and skills but lacks detailed trigger conditions. Environment fit: Primarily targets Claude Code environment, but does not specify compatibility with other environments.
Information architecture: README is well-structured with a table of contents and categories, but some links are broken. Install notes: Provides basic usage (e.g., running claude command) but lacks detailed installation steps. Naming stability: File names and command names are consistent, but no version history is provided. Examples and FAQ: Provides multiple examples and links, but no FAQ section. Known limitations: No known limitations mentioned. License: MIT license is clear. Versioning and changelog: No version numbers or changelog. Maintenance responsibility: README mentions the author but no explicit maintenance commitment.
Output usability: Documentation provides rich examples and best practices, but lacks actual output examples. Marginal value: As a best practices collection, it provides a lot of valuable information, but may overlap with other similar resources. Cost-benefit: Free and open source, but users need to assess its value themselves.
Claim traceability: README references multiple external links and documents, but some links may be broken. Cross-source corroboration: References multiple external sources but does not provide verification methods. Fact-inference separation: The documentation mixes facts and inferences without clear distinction.
- Not found in source: least-privilege scopingGrant only what the task needs: a dedicated account or read-only token, scoped to specific directories and repos.
- Not found in source: confirmation before actingTurn on (or add) a confirmation step before it acts, and try it in a sandbox or test environment before real data.
- Not found in source: data-flow disclosureWatch which external services it contacts (proxy or firewall logs) and keep sensitive data out until you know where it goes.
- Not found in source: sensitive-data handlingUse dedicated, low-privilege, revocable API keys — never production credentials — and keep secrets out of logs.
- Not found in source: dependency securityPin versions and run a dependency audit (npm audit, pip-audit) before installing; prefer running it in a container.
- Not found in source: disclosed external effectsEstablish which external systems it writes to, sends to or changes, and verify with test accounts or repos before production.
- Not found in source: rollback or recovery pathBack up first, or work on a git branch or snapshot, so its changes can be undone.
- Not found in source: verifiable attributionInstall from the official repo or registry and check the publisher and URL to avoid look-alike packages.
- The repository is primarily documentation and configuration, with no actual code, so its functionality or security cannot be verified.
- The README contains many broken links (e.g., '!/tags/...'), which may affect the usability of the documentation.
- No version numbers or changelog are provided, making it difficult to track updates.
- No known limitations are mentioned, so users may be unaware of potential issues.