AI Agents

13 plugins in this category

A structured methodology plugin that teaches Claude disciplined software workflows — from TDD cycles and systematic debugging to Socratic brainstorming and subagent-driven development.Superpowers teaches Claude how to work like a real software developer by giving it a toolkit of practical skills. You get methods for test-driven development, debugging that actually makes sense, brainstorming sessions, and even the ability to have other AI agents review code before it ships. The best part? You can build your own custom skills once you understand how the framework works.

tdd

Keeps your CLAUDE.md sharp — audits its contents, captures lessons from recent sessions, and ensures Claude's project memory stays accurate and useful.Your project's memory needs regular checkups too. This plugin helps you keep your CLAUDE.md file in good shape and actually useful—auditing what's there and capturing what you learn in each session so Claude has the right context when you need help. Better memory means Claude can work smarter with your codebase without constantly asking the same questions.

Guides a feature from idea to shipped code using dedicated agents for discovery, design decisions, and final review — a full structured workflow in one plugin.Building features without a plan is like renovating a house without blueprints. This plugin walks you through a seven-step process that starts with understanding what you're actually working with, nailing down what needs to happen, designing things properly, then implementing and reviewing your work. It keeps you from diving straight into coding and helps you catch issues before they become headaches.

Everything you need to build and iterate on Claude-powered agents using the official Agent SDK — designed to compress the full development loop.Build Claude Agent SDK applications with a development kit that handles everything from setting up your project to making sure it follows best practices. Whether you're working in TypeScript or Python, this kit takes care of the scaffolding work so you can focus on what makes your agent unique.

Write simple markdown rules that intercept specific Claude behaviors — a lightweight way to enforce guardrails and customize how Claude responds in your project.Hookify helps you set up safety guidelines for Claude using nothing but plain markdown. You can create rules that either warn you about or block risky stuff—like running dangerous commands, leaving debug code in production, or embedding passwords—and you don't need to touch a single line of code to do it.

hooks

guardrails

rules

Plugin Developer Toolkit

AI Agents 0

Seven expert skills covering every aspect of Claude Code plugin development — hooks, MCP servers, slash commands, agents, and validation best practices.Want to build Claude Code plugins? This toolkit gives you everything you need—seven specialized skills that cover hooks, MCP server integration, plugin structure, slash commands, agent development, and skill authoring. You'll get over a dozen working examples to learn from, six validation scripts ready for production use, and more than 21,000 words of documentation to guide you through every step.

Gives Claude a persistent memory across sessions — automatically extracts and compresses conversation history into tiered logs Claude can reference later.Keep Claude in the loop across your coding sessions. This plugin quietly saves what you've been working on, condenses the highlights through Claude Haiku each day, and brings those memories back when you return. No setup needed—Claude will know which files you touched, what you decided, and exactly where you left things hanging.

Build, refine, and benchmark custom Claude skills with built-in evaluation tooling — so you always know how well each skill is actually performing.Build Claude Code skills from scratch with Skill Creator, a toolkit that walks you through everything from your first idea to a polished, production-ready skill. You get four modes to work with—Create for building, Eval for testing, Improve for refining, and Benchmark for measuring performance—so you can develop and iterate on your skills with real structure and feedback at each step.

skills

claude-code

evaluation

Explanatory Output Style

AI Agents 0

Restores the explanatory tone — Claude explains the reasoning behind each implementation decision and highlights codebase patterns worth understanding.Need Claude to explain the *why* behind code suggestions, not just the *what*? This plugin brings back the old "Explanatory"" style by automatically adding educational context about implementation choices and codebase patterns to every response. Perfect if you learn better when someone walks you through the reasoning behind decisions rather than just handing you finished code."

Scans your codebase and recommends the specific hooks, skills, MCP servers, and subagents that will actually speed up your particular project.Claude Code Setup takes a look at your actual codebase and figures out which Claude Code automations would actually help you work faster. It digs into your project structure, what you're using, and how things are organized — then suggests specific integrations across MCP servers, skills, hooks, subagents, and slash commands that would make sense for your setup.

Scaffolds and validates AI agents using the Atomic Agents framework — specialized tooling and built-in best practice enforcement throughout the build process.Building AI agents doesn't have to be complicated. Atomic Agents gives you a complete workflow for creating intelligent agents from the ground up, with specialized helpers that guide you through schema design, planning out your architecture, reviewing code, and developing tools. It walks you through each step with clear guidance, teaches you skills as you need them, and keeps you on track with best practice validation so you actually build something solid.

agents

framework

architecture

Learning Output Style

AI Agents 0

Turns Claude into a coding mentor — it pauses at key decision points and asks for your input rather than just solving everything automatically.Turn your Claude Code sessions into a real learning experience instead of just watching solutions appear. This plugin puts you in the driver's seat by pausing at key moments and asking you to write 5-10 lines of code yourself—whether that's handling errors, building out business logic, or making architectural choices. You'll actually learn by doing instead of just reading along.

Keeps Claude in a focused loop — each pass it sees its own prior output and improves on it until the task is genuinely finished.Ralph Loop lets Claude keep working on a task across multiple attempts, automatically picking up where it left off each time. Instead of starting fresh, the plugin preserves all your file changes and git history between iterations, so Claude can learn from test failures and earlier attempts to improve the solution. It's like having an AI that can actually learn from its mistakes and refine its work without losing any progress.

iteration

loop

refinement

About Claude AI Agents Plugins

AI agent plugins extend Claude's capabilities beyond single-turn conversations, enabling it to plan, execute multi-step tasks, and coordinate with other AI systems. These plugins connect Claude to agent frameworks like AutoGPT, LangChain, and CrewAI, making it easy to build reliable automated workflows. Use them to delegate complex research, code generation pipelines, and data processing tasks to specialized sub-agents. Many plugins in this category also provide memory and state management so agents can maintain context across long-running tasks. Whether you're prototyping an agentic system or deploying production automation, these plugins give Claude the tools to act, not just respond.