Yggshanraisshan/claude-code-best-practice
↗ GitHubfrom vibe coding to agentic engineering - practice makes claude perfect
48,316
Stars
4,766
Forks
384
Watchers
9
Open Issues
Safety Rating A
The repository contains only documentation, markdown files, SVG/GIF assets, configuration examples, and curated links. No source code with executable logic, hardcoded secrets, suspicious dependencies, or obfuscated content was identified. The README contains no embedded instructions attempting to manipulate analyst behavior. This appears to be a legitimate, high-engagement community knowledge base.
ℹAI-assisted review, not a professional security audit.
AI Analysis
A comprehensive best-practices reference repository for Claude Code (Anthropic's AI coding agent), documenting concepts such as subagents, commands, skills, hooks, MCP servers, workflows, memory management, and settings. It aggregates tips, implementation examples, orchestration patterns, and curated resources (articles, videos, third-party workflow projects) sourced from the Claude Code team and community. The repository itself is partially maintained by Claude Code via automated workflows and serves as both a learning resource and a template for structuring Claude Code-powered projects.
Use Cases
- Learning Claude Code concepts including subagents, commands, skills, hooks, and MCP integrations
- Reference for best practices when building agentic engineering workflows with Claude Code
- Template for structuring .claude/ directories with agents, commands, skills, and settings
- Discovering and comparing community development workflow frameworks (BMAD-METHOD, Superpowers, etc.)
- Curating tips and tricks from the Claude Code creator and team for daily productivity
Tags
Project Connections
claude-scientific-skills
→Claude Scientific Skills follows the same Agent Skills standard (.claude/skills/) documented in this best-practices repo, and this repo's guidance on skills, subagents, and commands directly applies to how Claude Scientific Skills are structured and consumed.
zeroleaks
→The best-practices repo documents Claude Code's security-related features (prompt injection defense, permission modes, Auto Mode safety classifier) and zeroleaks provides adversarial testing of LLM systems, making them complementary tools for securing agentic Claude Code deployments.
free-code
→free-code is a fork of Claude Code CLI with guardrails removed, targeting the same Claude Code user base but taking an opposing philosophy to the safety and best-practice guidance promoted in this repository.
marketing-dashboard
→The marketing-dashboard integrates with agent runtimes and documents agent/squad discovery patterns that align with the orchestration workflow (Command → Agent → Skill) patterns described in this best-practices repo.
CoPaw
→CoPaw implements multi-agent collaboration, MCP client management, scheduled tasks, and memory — all topics extensively covered in this best-practices repo — making it a practical complement for teams adopting Claude Code patterns in a broader multi-platform agent framework.