| agent-sdk-dev | A software development kit that enables developers to build applications and integrations using Anthropic's Claude Agent SDK. | Development | Apache-2.0 | Anthropic |
| ai-plugins | Endor Labs is a security tool that installs via endorctl to analyze software supply chains, identifying, ranking, and remediating vulnerabilities and related risks. | — | MIT | — |
| aikido | Aikido Security integrates with Claude Code via its MCP server to perform static application security testing, secrets detection, and infrastructure-as-code vulnerability scanning. | — | MIT | — |
| altimate-code | Assigns dbt and data warehouse tasks to altimate-code, a dedicated CLI agent offering 100+ specialized tools for SQL analysis, lineage tracking, dbt operations, warehouse profiling, cost management, and connections to major database platforms. | Database | MIT | AltimateAI |
| amd-skills | An AMD plugin bundling verified Agent Skills: routes image/audio processing to local AI on Ryzen AI, deploys LLMs via vLLM on AMD Instinct GPUs, boosts inference throughput using Hyperloom, and analyzes GPU kernel and PyTorch trace performance. | Development | MIT | AMD |
| apollo-skills | Apollo-skills is a plugin bundling reference guides and instructions for AI coding agents (e.g., Claude Code), covering Apollo GraphQL tools like Client, Server, Federation, Router, Rover CLI, and MCP Server, installable via CLI or as a plugin. | Development | MIT | Apollo GraphQL |
| atomic-agents | A structured toolkit supporting AI agent development within the Atomic Agents framework, offering specialized agents for schema, architecture, code review, and tool creation, plus guided workflows and best-practice checks. | Development | MIT | — |
| azure-cosmos-db-assistant | A Claude Code plugin adding Azure Cosmos DB expertise—data modeling, query optimization, partition keys, SDK usage—via 73 categorized rules, an expert agent, and review/regeneration commands; optionally connects to live databases via MCP Toolkit. | Database | MIT | — |
| base44 | A CLI and SDK toolkit for creating, managing, and deploying full-stack applications on Base44, supporting project setup and JavaScript/TypeScript development workflows. | Development | MIT | — |
| buildkite | Official Buildkite integration skills enabling AI coding agents to manage pipelines, migrations, preflight checks, agent runtime configuration, CLI usage, and API interactions. | Development | MIT | Buildkite |
| carbone-skill | Official Carbone documentation detailing its template syntax—tags, loops, conditionals, formatters, and aggregators—usable when generating documents in office, web, or portable file formats. | Productivity | Apache-2.0 | Carbone |
| claude-code-setup | Claude Code plugin that analyzes a project's codebase to recommend suitable automation features—MCP server integrations, skills, hooks, subagents, and slash commands—without modifying files. Created by Isabella He. | Productivity | Apache-2.0 | Anthropic |
| claude-md-management | A toolset for maintaining and improving CLAUDE.md files by checking their quality, recording insights from work sessions, and keeping project documentation up to date. | Productivity | Apache-2.0 | Anthropic |
| claude-security | Performs in-depth code vulnerability scanning within Claude Code sessions at adjustable effort levels, verifying findings via code-based tallying and generating agent-panel-checked patches for optional application. | Security | Proprietary | Anthropic |
| code-modernization | This tool guides legacy software modernization: running /modernize produces an assessment, interactive map, extracted business rules, and an approvable plan, followed by upgrading, rewriting, or rebuilding code with verified behavioral equivalence, security improvements, and reporting. | Development | Apache-2.0 | Anthropic |
| code-review | This tool automates pull request code reviews by deploying multiple specialized agents that assign confidence scores to findings, reducing false positive results. | Productivity | Apache-2.0 | Anthropic |
| code-simplifier | This tool refactors recently changed code to improve clarity, consistency, and maintainability without altering its original functionality. | Productivity | Apache-2.0 | Anthropic |
| commit-commands | Provides commands to streamline git workflows, covering committing changes, pushing to remote repositories, and creating pull requests. | Productivity | Apache-2.0 | Anthropic |
| confidence | This tool integrates Confidence's feature flag, experiment, and migration management capabilities into Claude Code, allowing direct access to these functions within that environment. | Development | Apache-2.0 | Spotify Confidence |
| convex | Official Convex plugin for Claude Code offering backend skills, a code-writing subagent, runtime-error monitoring, and MCP access for schema, real-time, auth, storage, jobs, and AI agents. | Database | Apache-2.0 | Convex |
| cwc-makers | Sets up a Code-with-Claude Makers Cardputer via a single /maker-setup command, cloning the repo, flashing UIFlow firmware, and installing the Claude Buddy app. | Productivity | Apache-2.0 | Anthropic |
| dash0 | Instruments Claude Code sessions with OpenTelemetry, recording tool calls, LLM requests, token consumption, and errors as traces sent to Dash0 or other OTel-compatible backends. | Monitoring | Apache-2.0 | Dash0 |
| desktop-commander | A Claude Code plugin combining an MCP server with skills for terminal use, system checks, AI setup, and knowledge management, plus a paid desktop app enabling local, natural-language task automation across apps like HubSpot, Slack, Linear, and WordPress. | Productivity | MIT | Desktop Commander |
| discord | Discord messaging bridge featuring built-in access control, allowing management of pairing, allowlists, and policies through /discord:access. | Productivity | Apache-2.0 | — |
| duckdb-skills | A set of DuckDB-based Claude Code capabilities for reading various data files, connecting to and querying DuckDB databases, searching DuckDB/DuckLake documentation and prior session logs, and managing DuckDB extension installation and updates. | Database | MIT | DuckDB Foundation |
| duende-skills | A Claude Code extension providing development skills and agents for Duende, addressing OAuth/OIDC, IdentityServer, token handling, ASP.NET Core auth, BFF patterns, and secure identity system design. | Security | MIT | Duende Software |
| explanatory-output-style | This tool provides educational commentary on implementation decisions and coding patterns within a codebase, recreating the functionality of the discontinued Explanatory output style. | Learning | Apache-2.0 | Anthropic |
| fakechat | A local test chat tool for simulating channel notification flows on localhost, requiring no tokens, access control, or third-party services. | Development | Apache-2.0 | — |
| fastly-agent-toolkit | Fastly platform tools for configuring and managing CDN, edge computing, and web performance services, aimed at developers building or optimizing Fastly-based infrastructure. | — | MIT | — |
| feature-dev | This tool coordinates specialized agents to explore codebases, design architecture, and review quality throughout the software feature development process. | Development | Apache-2.0 | Anthropic |
| gitkraken | Provides Claude with access to actual Git and project data—commits, branches, pull requests, and issues—across repositories, supporting GitHub, GitLab, Azure DevOps, Bitbucket, and Jira. | Development | MIT | GitKraken |
| growthbook | A collection of agent skills covering the entire lifecycle of GrowthBook feature flags and experiments, from setup through management and analysis. | Testing | MIT | GrowthBook |
| honeycomb | A collection of skills, agents, and workflows supporting Honeycomb observability tasks, including query patterns, production troubleshooting, SLO management, OpenTelemetry instrumentation, and migrating from Beeline, built to work alongside the Honeycomb MCP server. | Monitoring | MIT | Honeycomb |
| hookify | This tool generates custom hooks that block unwanted behaviors by examining conversation patterns or explicit user instructions, with rules configured in plain markdown files. | Productivity | Apache-2.0 | Anthropic |
| huggingface-skills | Enables users to build, train, evaluate, and deploy open source AI models, along with accessing datasets and shared project spaces. | Development | Apache-2.0 | — |
| idmp-plugin | A plugin for TDengine IDMP that provides built-in capabilities for discovering resources, inspecting schemas, and executing operational workflows safely. | Development | MIT | TaosData |
| imessage | This tool bridges iMessage with permission management, reading chat.db for messages and using AppleScript to send them, controlling access through pairing and allowlists. | Productivity | Apache-2.0 | — |
| jfrog | This tool integrates the JFrog Platform with Claude Code, enabling access to Artifactory repositories, security findings, package safety data, SDLC workflows, and platform administration tasks. | Security | Apache-2.0 | JFrog Ltd. |
| langfuse-observability | A plugin that captures Claude Code's traces, spans, and session telemetry, exporting them to Langfuse to support LLM monitoring, debugging, and evaluation. | Monitoring | MIT | Langfuse |
| learning-output-style | This tool is an interactive learning mode prompting users for meaningful code input at key decisions, replicating the discontinued Learning output style. | Learning | Apache-2.0 | Anthropic |
| liquid-skills | This resource covers Liquid templating basics, standard practices for CSS/JS/HTML coding, and WCAG accessibility guidelines applied to Shopify theme development. | Development | MIT | Shopify |
| lovable | A Claude Code extension integrating the Lovable MCP server via OAuth 2.1, providing commands to build, iterate, deploy, and manage Lovable apps and databases, with safety prompts for credits and publishing. | Development | Apache-2.0 | Lovable |
| math-olympiad | This tool solves competition-level math problems (IMO, Putnam, USAMO) using adversarial verification, where independent verifiers probe proofs for specific flaws, favoring calibrated abstention over false confidence. | Math | Apache-2.0 | Anthropic |
| math-proof | Two approaches for tackling difficult math problems, both producing a standalone proof.md: one has a single session solve it progressively using notes; the other runs prolonged rounds of judge and worker sub-agents, clearly distinguishing proven from unproven claims. | Math | Apache-2.0 | Anthropic |
| mattpocock-skills | A collection of practical agent skills by Matt Pocock covering software engineering practices like requirement clarification, spec/ticket workflows, test-driven development, code review, and domain modeling, designed for ready-to-use application rather than improvised coding. | Development | MIT | Matt Pocock |
| mcp-apps | This tool provides skills for building MCP Apps, leveraging the MCP Apps SDK to guide developers through application creation. | Development | MIT | Anthropic / Model Context Protocol |
| mcp-server-dev | Provides guidance for creating MCP servers compatible with Claude, covering deployment options (remote HTTP, MCPB, local), tool design approaches, authentication methods, and building interactive MCP applications. | Development | Apache-2.0 | Anthropic |
| mcp-tunnels | This command sets up a private MCP server tunnel to Claude via Docker Compose, handling certificates, proxy configuration, cloudflared, and a test server automatically. | Development | Apache-2.0 | Anthropic |
| mintlify | Mintlify is a documentation site builder that converts non-markdown files into formatted MDX pages, supports adding and editing content with proper component usage, and automates documentation updates. | Development | MIT | — |
| modern-web-guidance | A tool that helps keep coding agents current with up-to-date web development best practices, ensuring generated code follows modern standards. | Development | Apache-2.0 | Google Chrome |
| netlify-skills | Set of Claude Code capabilities covering Netlify's platform: managing functions, edge functions, blobs, databases, image CDN, forms, configuration, CLI usage, framework support, caching, AI gateway integration, and deployment tasks. | Development | MIT | — |
| newrelic | A Claude Code integration that connects to New Relic telemetry, enabling terminal-based APM performance investigation, cloud cost analysis, Kubernetes debugging, NRQL query writing, and alert response. | Monitoring | Apache-2.0 | New Relic |
| nvidia-skills | A discovery tool that helps users locate suitable NVIDIA-related skills covering GPU acceleration, CUDA, AI agents, data handling, model training and inference, robotics, Physical AI, Omniverse, and simulation tasks. | Development | Apache-2.0 AND CC-BY-4.0 | NVIDIA |
| outputai | Output.ai's toolkit extends Claude Code with five specialized agents (planning, building, debugging, prompt writing, quality review), over 40 slash-command skills for scaffolding, debugging, evaluation, and credential handling, and a SessionStart hook preloading Output SDK conventions. | Development | Apache-2.0 | Output.ai |
| pagerduty | Analyzes code changes by comparing them to historical incident data, generating PagerDuty-based risk scores to flag deployment risks before shipping. | Monitoring | Apache-2.0 | — |
| paypal | Adds PayPal integration to Claude Code, providing guidance skills, slash commands, and a pre-write hook, plus an MCP server exposing tools for orders, invoices, subscriptions, disputes, and reporting for developers. | Development | Apache-2.0 | PayPal |
| pixeltable | Pixeltable is a Python library for building multimodal AI applications, offering tables with computed columns, embedding search, UDF support, tool-calling agents, and integrations with over 25 AI providers. | Development | Apache-2.0 | Pixeltable |
| playground | Generates interactive, single-file HTML playgrounds featuring visual controls, live previews, and copyable prompt output, with templates for design, data, concepts, and document review. | Development | Apache-2.0 | Anthropic |
| plugin-dev | A toolkit for building Claude Code plugins, offering seven expert skills on hooks, MCP integration, commands, and agents, plus AI-assisted creation and validation. | Development | Apache-2.0 | Anthropic |
| pr-review-toolkit | A suite of automated agents that perform thorough pull request reviews, covering comment quality, test coverage, error handling, type design, overall code quality, and simplification opportunities. | Productivity | Apache-2.0 | Anthropic |
| preset-cli-skills | Provides preset CLI skills for shell, scripting, and CI/CD tasks via the sup command-line tool, intended solely for CLI-based workflows. | Development | Apache-2.0 | Preset |
| project-artifact | This tool creates and updates a shareable claude.ai status page for a project, showing overview, goals, workflow stages, and next steps, refreshing data while keeping the same URL and reporting only changes. | Productivity | Apache-2.0 | Anthropic |
| pydantic-ai | Provides guidance for writing correct Pydantic AI code, covering current patterns, decision logic, and common pitfalls across agents, tools, structured outputs, streaming, and multi-agent systems. | Development | MIT | — |
| qdrant-skills | A collection of agent skills addressing Qdrant vector database operations—scaling, performance tuning, search quality, monitoring, deployment, model migration, and version upgrades—plus SDK guidance for Python, TypeScript, Rust, Go, .NET, and Java. | Database | Apache-2.0 | Qdrant |
| qodo | Qodo integrates codebase intelligence, coding standards enforcement, and code review capabilities directly into a coding agent, helping ensure consistent, high-quality code development within that agent's workflow. | Development | MIT | Qodo |
| qodo-standards | Qodo offers configurable workflows for discovering optional coding rules and managing organizational standards, helping teams locate applicable guidelines and administer compliance practices within their development processes. | Development | MIT | Qodo |
| qt-development-skills | A skill set for Qt software engineering tasks, covering reviewing Qt C++/QML code, writing QML code, and producing documentation for Qt C++/QML codebases. | Development | Proprietary | Qt Group |
| ralph-loop | An AI development tool that repeatedly runs Claude on a task, allowing it to review prior iterations and refine work until the task is complete. | Development | Apache-2.0 | Anthropic |
| rc | A plugin toolkit enabling AI coding assistants like Claude Code to manage RevenueCat project setup—apps, products, entitlements, offerings—via MCP server and platform-specific integration skills, with OAuth account authentication. | Development | MIT | — |
| receipts | Generates a personal usage report for Claude Code, summarizing shipped work, affected projects, and usage distribution by analyzing local transcripts and read-only git data; only counts/names are transmitted, with the report kept locally. | Productivity | Apache-2.0 | Anthropic |
| redis-development | A guide covering recommended approaches for developing with Redis, addressing data structures, its query engine, vector search capabilities, caching strategies, and techniques for optimizing performance. | Database | MIT | Redis |
| remember | A memory system for Claude Code that captures conversation history, condensing it into layered daily logs so Claude retains context from previous sessions. | — | Proprietary | — |
| render | A Render deployment toolkit offering deployment, debugging, and monitoring capabilities, featuring built-in skills, an agent component, slash commands, and a hook for validating render.yaml configuration files. | Deployment | MIT | Render |
| rill | A toolkit for building and interacting with Rill business intelligence projects, enabling development work and data querying within the Rill platform. | Development | Apache-2.0 | Rill Data |
| rootly | Provides end-to-end incident management covering deployment safety, incident response coordination, on-call scheduling, and post-incident retrospectives. | Monitoring | Apache-2.0 | Rootly |
| save-to-spotify | This tool generates audio episodes using TTS narration combined with timelines and cover images, then uploads the resulting content to Spotify via a CLI utility. | Productivity | Apache-2.0 | Spotify |
| security-guidance | This tool checks Claude-generated code for security issues, combining pattern-based edit warnings, LLM diff review, and commit analysis to detect injection, XSS, SSRF, hardcoded secrets, and other vulnerabilities. | Security | Apache-2.0 | Anthropic |
| semgrep | Semgrep detects security vulnerabilities as code is written, helping guide Claude toward producing secure code from the outset. | Security | — | — |
| sentry | Sentry integration for retrieving error reports and stack traces, searching issues by fingerprint, enabling developers to debug production errors from within their development environment. | Monitoring | MIT | — |
| session-report | Creates an interactive HTML report analyzing Claude Code session data from local transcripts, covering token usage, cache efficiency, subagents, skills, and costliest prompts. | Productivity | Apache-2.0 | Anthropic |
| skill-creator | This tool supports creating new skills, refining existing ones, and evaluating their performance through testing and variance-based benchmarking analysis. | Development | Apache-2.0 | Anthropic |
| snowflake-cortex-code | Routes Claude Code's Snowflake-related prompts to Cortex Code for execution, offering slash commands for code review and delegation, along with routing, run, and setup skills. | Development | Apache-2.0 | Snowflake |
| sonarqube | Integrates SonarQube into AI coding agents, automatically checking code quality, security, and secrets across 40+ languages using hooks, CLI, MCP server, and slash commands. | Security | Proprietary | SonarSource |
| sourcegraph | This tool enables searching, reading, and tracing code references across repositories, assessing refactoring impact, investigating incidents through commit/diff history, and conducting security-focused code scans. | Development | MIT | — |
| stripe | This tool integrates Stripe development capabilities into Claude, enabling assistance with Stripe-related coding tasks such as API integration and payment processing. | Development | MIT | — |
| superpowers | Superpowers is a training resource enabling Claude to brainstorm, use subagent-driven development with code review, debug systematically, apply red/green TDD, and create and test new skills. | Development | MIT | — |
| teamcity-cli | This is a Claude agent skill that uses the teamcity CLI to interact with TeamCity CI/CD, managing builds, logs, queues, and agents. | Development | Apache-2.0 | JetBrains |
| telegram | A messaging bridge that connects to Telegram, featuring integrated access control for managing pairing, allowlists, and policy through the /telegram:access command. | Productivity | Apache-2.0 | — |
| togetherai-skills | Provides Agent Skills for Together AI, enabling inference, training, embeddings, and media generation via serverless chat, fine-tuning, batch processing, evaluations, and GPU infrastructure. | Development | MIT | Together AI |
| unity | Unity's official Claude Code plugin offering built-in skills focused on game development tasks, monetization strategies, and performance optimization for Unity projects. | Development | Proprietary | Unity Technologies |
| unreal-engine-skills-for-claude-code | This tool connects Claude Code to Unreal Editor via MCP, enabling control through hundreds of ToolsetRegistry functions spanning 30+ areas like actors, blueprints, materials, Niagara, Sequencer, and gameplay systems. | Development | MIT | Epic Games |
| vanta | Vanta plugin links Claude Code with Vanta's compliance platform via Vanta MCP server, using remediation guidance and local repo context to speed up fixing compliance test failures. | Security | MIT | Vanta |
| vanta-mcp-plugin | Vanta plugin links Claude Code to Vanta's compliance platform via Vanta MCP server, combining remediation guidance with local repo context to fix compliance failures. | Security | MIT | Vanta |
| vercel | Enables managing Vercel deployments through Claude Code, including tracking build status, viewing logs, configuring domains, and controlling frontend infrastructure directly from the interface. | Deployment | Apache-2.0 | — |