The best plugins and skills for AI coding agents

A battle-tested collection of skills and workflows that gives Claude Code structured planning, test-driven development habits, and self-improvement loops out of the box.

Forces the laziest solution that actually works: standard library before custom code, native platform features before dependencies, one line before fifty. Kills over-engineering on sight.

Behavioral guidelines distilled from Andrej Karpathy's advice for LLM coding: surgical changes, surfaced assumptions, verifiable success criteria, and no unrequested complexity.

Official Anthropic skills for document work: create and edit Word, PowerPoint, Excel, and PDF files with proper formatting - the same skills that power Claude document handling.

CLI with hundreds of ready-made agents, commands, hooks, and MCP configurations for Claude Code - browse, pick, and install per project or globally.

A large collection of focused Claude Code subagents - code reviewers, security auditors, database experts, language specialists - each tuned for one production task.

Curated index of the best Claude Code commands, workflows, hooks, and plugins from across the ecosystem - the fastest way to survey what exists.

Categorized collection of specialized subagent definitions - full-stack, DevOps, data, ML - ready to drop into .claude/agents for multi-agent workflows.

General-purpose biomedical AI agent environment from Stanford: hundreds of biology tools, databases, and protocols an LLM can drive for research tasks like CRISPR screening and drug repurposing.

Grammarly-style Enhance button for prompts on ChatGPT, Claude, Grok, and Gemini - plus an MCP server that brings the same rewriter into Cursor, Claude Desktop, and VS Code.