GitRead 中文社区
首页趋势排行榜原创文章严选周刊
更多资源
新手训练营
精选专题热门话题GitHub 文档资源精选实战指南

awesome-claude-code-plugins

精选 Claude Code 插件——Claude Code 的斜杠命令、子代理、MCP 服务器和钩子精选列表

  • 语言Python
  • 默认分支main
  • 大小2.1 MB
  • 更新8月12日
  • Issues493
在 GitHub 打开
avatarccplugins

技术解读

以下内容由 GitRead 中文社区 编辑团队基于项目 README 与官方资料整理, 结合中文开发者的常见使用场景补充技术选型建议。具体以项目官方文档为准。

把提示词变成可分发资产:awesome-claude-code-plugins 的策展逻辑与工程边界

Claude Code 的扩展能力一直很像"散落的插件市场":官方文档告诉你 settings.json 里可以挂 Hook,博客告诉你怎么写 .claude/commands/,Twitter 上流传一段 review.md,你抄下来改两个词,然后它就悄悄在你的仓库里执行 git push。ccplugins/awesome-claude-code-plugins 要处理的正是这种"能力存在、但无法发现、无法复用、无法信任"的状态。

项目背景与痛点

Claude Code 的扩展面有四个入口:斜杠命令、子代理(Subagent)、MCP Server、Hook。它们的共同点是用文件做配置、用 Markdown 做代码——门槛极低,但代价是缺少类型、缺少版本、缺少语义约束。生态因此迅速碎片化:同一个"代码审查"命令有几十个变体,质量取决于作者当天的心情;MCP Server 的配置格式各写各的;Hook 脚本更像是从某篇推文里粘来的黑盒。

这个仓库的定位不是框架,而是索引层:按扩展类型分层归类,给出可直接落地的仓库地址与安装路径(含 Plugin Marketplace 的 /plugin marketplace add 流程)。它解决的是"发现与规范化"问题,而不是"运行时"问题——这一点必须看清,否则会误判它的价值。

核心技术揭秘

要理解这个列表的组织方式,得先还原 Claude Code 的扩展数据流。四类扩展点并非平级:命令是提示词的宏展开,子代理是上下文的隔离与委派,MCP 是唯一有协议规范的外部能力,Hook 则是围绕工具调用的拦截器。

graph TD A[用户输入 Prompt] --> B{"扩展点路由"} B -->|"斜杠命令"| C[Slash Command 模板展开] B -->|"子代理委派"| D[Subagent 隔离上下文] B -->|"外部能力"| E[MCP Server 工具调用] C --> F[主 Agent 推理循环] D --> F E --> F F --> G[PreToolUse Hook 校验] G -->|"放行"| H[执行工具调用] G -->|"拦截"| I[中止并回传原因] H --> J[PostToolUse Hook 记录审计] J --> K[汇总输出]

从这张图能看出几个工程事实:

  • 命令与子代理本质上是"提示词资产"。命令依赖 frontmatter 里的 description、argument-hint、allowed-tools 做约束,正文支持 $ARGUMENTS、@file 引用和 ! 内联 shell 的输出注入;子代理则通过 tools 白名单和独立上下文窗口换取"不污染主会话"的收益,只把结论回传。
  • Hook 是唯一的强干预点。它在 settings.json 中按事件注册(PreToolUse、PostToolUse、UserPromptSubmit、Stop 等),通过 stdin 接收 JSON、以退出码表达放行或阻断。这意味着它天然适合做审计与合规,也天然是最大的风险面。
  • MCP 是列表里唯一可以谈"协议"的部分。stdio / SSE / HTTP 三种传输、工具以 mcp__server__tool 形式暴露,这是唯一具备跨厂商复用潜力的层。

因此该仓库的分类逻辑其实是对的:把"提示词片段"和"可执行组件"分开。它的技术难点不在代码,而在策展的熵控——去重、失效链接治理、质量分级,以及最容易被忽视的:同一类目下命令的语义冲突(两个都叫 /review 的插件装在一起会怎样)。Plugin Marketplace 机制部分缓解了这个问题,因为插件可以整包携带 commands、agents、hooks 与 MCP 配置,安装即约定,但版本锁定与冲突检测依然靠人。

功能亮点与差异

相比零散的 gist 合集,它的杀手锏是按扩展点建模的分类法,而不是单纯的链接堆叠。这让开发者可以按"我需要的是流程编排还是工具接入"来检索,而不是按热度盲选。

另外值得肯定的是它把 Plugin 作为分发单元的思路:一个插件包内聚了命令、代理、Hook 和 MCP 依赖,安装路径统一,比"复制一个 md 文件到 .claude/commands/"更接近可运维的形态。组合模式也比单点更强——命令负责入口,子代理负责重活,Hook 负责收尾审计,这是真正能拼出工作流的用法。

应用场景与落地建议

适合的场景很明确:团队内标准化工作流(统一的提交信息、审查清单、发布前置检查)、把内部系统接成只读工具(知识库、工单查询走 MCP)、在 CI 或本地做工具调用的审计留痕。

生产环境使用必须守住几条线:

  1. 最小权限。命令的 allowed-tools、子代理的 tools 白名单能收就收,不要默认放开 shell。
  2. Hook 是代码,不是配置。它执行任意命令,务必审查脚本内容、处理非零退出码语义,避免一次误判把整条流水线卡死。
  3. MCP Server 属于供应链。凭证走环境变量,优先本地 stdio,远程端点要评估数据外泄面。
  4. 固定版本。不要把 main 分支直接接入关键流程;插件一旦更新,提示词行为可能整体漂移。
  5. 防递归委派。子代理再调子代理会放大 token 消耗与时延,用工具白名单切断。

综合评价

优点:分类维度贴着 Claude Code 的真实扩展模型,不是无脑聚合;把插件作为分发单元,落地路径清晰;对"提示词即资产"这一趋势的判断是准确的。

缺点:本质仍是 awesome list,质量不可验证、缺少测试样例与兼容性标注;更新节奏依赖社区,链接腐化难以避免;最关键的是安全语义缺失——Hook 与 MCP 的风险等级没有在列表里体现,而这恰恰是生产采纳的第一道门槛。

一句话:它是一张不错的地图,但不是护栏。用它找方向,用工程规范兜底。

内容已同步仓库最新状态(最后提交:8月12日)。 编辑团队负责选题与抽样审核,详见 编辑准则。

相关仓库

public-apis/public-apis avatar
public-apispublic-apis
1 天前
497.4k
55.6k
2.6k

免费 API 汇总列表

apipublic-apisfree+7
Python
15.5 MB
7
详情
EbookFoundation/free-programming-books avatar
free-programming-booksEbookFoundation
1 天前
401.3k
67.5k
95

免费可获取的编程书籍

bookseducationhacktoberfest+2
Python
21.8 MB
3
详情
donnemartin/system-design-primer avatar
system-design-primerdonnemartin
9月15日
377.4k
57.4k
290

学习如何设计大规模系统。为系统设计面试做准备。包含 Anki 闪卡。

designdesign-patternsdesign-system+10
Python
11.0 MB
0
详情
vinta/awesome-python avatar
awesome-pythonvinta
21 小时前
333.5k
29.1k
19

回答“我想用 Python 做 X,该用哪个工具?”的权威清单

awesomeawesome-listpython+3
Python
6.9 MB
2
详情
practical-tutorials/project-based-learning avatar
project-based-learningpractical-tutorials
5 小时前
289.4k
36.8k
238

精选的项目实战教程列表

tutorialprojectbeginner-project+5
Python
472 KB
1
详情
NousResearch/hermes-agent avatar
hermes-agentNousResearch
1 小时前
261.9k
58.3k
53.8k

与你共同成长的智能体

aiai-agentai-agents+10
Python
1.4 GB
3
详情
TheAlgorithms/Python avatar
PythonTheAlgorithms
6 天前
226.0k
51.3k
90

所有算法均用 Python 实现

algorithmalgorithm-competitionsalgorithms-implemented+11
Python
17.9 MB
2
详情
yt-dlp/yt-dlp avatar
yt-dlpyt-dlp
5 天前
206.4k
18.1k
2.8k

功能丰富的命令行音视频下载器

youtube-dlpythonsponsorblock+4
Python
63.5 MB
0
详情

关于

claude-codeclaude-code-commandsclaude-code-hooksclaude-code-mcpclaude-code-pluginclaude-code-subagentsawesome-claude-codeawesome-listawesome-resourcesawesome-plugins
  • Apache License 2.0
  • 1.0k Stars
  • 832 Forks
  • 9 Watching

简要分析

热度分
2673
价值分
825
活跃状态
活跃
GitHub

语言占比

Batchfile
Dockerfile
JavaScript
Python
Shell

贡献者

dudley02dudley02hanhuarkhanhuarkHiH-DimaNHiH-DimaNBayramAnnakovBayramAnnakovdotcomjackdotcomjackabullard1abullard1dybovalexdybovalexhuiyu9144huiyu9144piiiicopiiiicoIbrahim-3dIbrahim-3dununununiumununununiumxiaolaixiaolai

Awesome Claude Code Plugins Awesome

English 简体中文

Awesome Claude Code plugins — a curated list of slash commands, subagents, MCP servers, and hooks for Claude Code

  • What is Claude Code Plugin?
  • Use Cases
  • Plugins
    • Official Claude Code Plugins
    • Workflow Orchestration
    • AI & Speech
    • Automation DevOps
    • Business Sales
    • Code Quality Testing
    • Communication & Integrations
    • Data Analytics
    • Audio & Media
    • Design UX
    • Development Engineering
    • Documentation
    • Git Workflow
    • Marketing Growth
    • Project & Product Management
    • Lifestyle & Entertainment
    • Security, Compliance, & Legal
    • MCP Servers
    • Thinking & Knowledge Management
    • Companion Apps & Tools
    • Knowledge Management
  • External Marketplaces
    • Skills & Frameworks
  • Resources
  • Tutorials
  • Marketplaces
  • Contributing

What is Claude Code Plugin?

Claude Code Plugin is lightweight package that let you customize and share your Claude Code setup. Each plugin can include any combination of:

  • Slash Commands — Custom shortcuts for frequent operations
  • Subagents — Purpose-built agents for specialized dev tasks
  • MCP Servers — Integrations to tools and data sources via the Model Context Protocol
  • Hooks — Extensions that modify Claude Code’s behavior at key workflow points

Install or disable them dynamically with the /plugin command — enabling you to keep your system context focused and lightweight.

Use Cases

  • Enforce Standards: Ensure specific hooks or workflows run consistently across your team
  • Support Users: Package slash commands that simplify your framework or SDK usage
  • Share Workflows: Publish debugging setups, deployment scripts, or testing harnesses
  • Connect Tools: Integrate internal systems securely through MCP servers
  • Bundle Customizations: Combine multiple extensions for a cohesive developer experience

Plugins

Official Claude Code Plugins

  • agent-sdk-dev
  • pr-review-toolkit
  • commit-commands
  • feature-dev
  • security-guidance

Workflow Orchestration

  • agent-triforce - 3-agent PM/Dev/QA system with WHO checklist methodology. Prometeo (PM), Forja (Dev), Centinela (QA) with 24 checklists, 6 skills, and auto-generated HTML dashboard.
  • angelos-symbo
  • artel - Infrastructure for AI teams: shared semantic memory, tasks, agent-to-agent messages, and session handoffs across machines and LLM providers.
  • claude-brain — Sync and evolve your Claude Code brain across machines. Auto-syncs memory, skills, agents, rules, and CLAUDE.md via Git with LLM-powered semantic merge.
  • ceo-quality-controller-agent
  • claude-recap — Per-topic session memory using Shell hooks — archives each conversation topic as a separate Markdown summary. Two hooks, bash + Node.js, 100% local.
  • claude-desktop-extension
  • equilateral-agents - 22 self-learning agents with memory, security review, code quality, deployment validation, and infrastructure checks
  • lyra
  • magebyte-power — Production-incident-distilled Claude Code skill: 7-phase workflow with 4-round AI cross-verification that catches concurrency & idempotency bugs before prod
  • model-context-protocol-mcp-expert
  • now-next-methodology - Two-file task management system for Claude Code/OpenCode: NOW.md tracks current work, NEXT.md queues future tasks with /next command
  • problem-solver-specialist
  • reviewer-author-loop
  • studio-coach
  • tree-ring-memory
  • ultrathink
  • pro-workflow - Self-correcting memory, 17 skills, 6 agents, 16 commands, 23 hooks. Context engineering, parallel worktrees, and agent teams.
  • claude-reflect - Self-learning system that captures corrections during sessions and syncs to CLAUDE.md with human review
  • agent-manager-skill
  • conductor-orchestrator-superpowers
  • context-mode
  • Claude Forge — oh-my-zsh for Claude Code: 11 agents, 36 commands, 15 skills, 14 hooks, 6-layer security. Install in 5 min.
  • claude-memory-manager - Global memory tier for Claude Code. Stores cross-project memories in ~/.claude/memory/, injects them via SessionStart hook. Includes a web UI for managing all memory files without digging through directories.
  • ultraship
  • launch-swarm - 6-agent Claude Code swarm with strict role separation, dual-gate merge pipeline, and prompt health anti-rot
  • cc-inspect
  • omega-memory
  • idea-to-deploy
  • nexus-agents — Multi-CLI orchestration with 30 MCP tools, 12 expert agents, 17 skills, and Budget→TOPSIS→LinUCB routing. Coordinates Claude, Gemini, Codex, and OpenCode via consensus voting.
  • bobusang — Multi-device memory system for Claude Code. Syncs context across Windows, WSL2, and Linux with git-based auto-sync.
  • magic-cc-codex-worker
  • weft
  • hyperflow
  • espresso - Full token-saving stack. Output compression, global rules, RTK, Caveman, GitNexus auto-config. Detects existing setup, installs only what’s missing.
  • aicoo-skills - AI COO agent skills: share agent via links, agent-to-agent communication, autonomous heartbeat loops, context sync, Aicoo Square discovery board
  • devforge-ai — Agentic SDLC orchestration for Claude Code: 10 role-specific agents across Plan → Build → Verify → Ship → Operate phases with tracer bullet delivery
  • ai-meeting
  • praetor
  • fable-baton
  • craftsman
  • rote
  • claude-session-tint

AI & Speech

  • speech-ai - Speech AI plugin with pronunciation assessment, text-to-speech, and speech-to-text. 8 MCP tools for language learning, accessibility, and voice applications.

Automation DevOps

  • PUIUX Pilot - Auto-configures Claude Code hooks, MCPs, and skills for any project. Scans 95+ project types, selects from 28+ hooks, scores quality (0-100, A-F). npm i -g puiux-pilot
  • deployment-engineer
  • devops-automator
  • MyVibe - Instant deployment with /myvibe:publish
  • infrastructure-maintainer
  • kubestellar-console
  • monitoring-observability-specialist
  • n8n-workflow-builder
  • qovery-deploy - Deploy any app to Kubernetes on AWS/GCP/Azure/Scaleway. Creates Dockerfiles, provisions databases, deploys via CLI+API or Terraform. Install: curl -fsSL https://skill.qovery.com/install.sh | bash
  • screenshot-janitor
  • aws-cost-saver
  • prompt-chime - Cross-platform notification plugin with meme sound presets, toast notifications, profiles, quiet hours, and per-project overrides.
  • cc-aws-keepalive
  • claude-snapshot - Portable .tar.gz snapshots of your Claude Code setup (settings, hooks, plugins, MCPs) with diff-before-apply and .bak rollback for migration and backup across machines.
  • notify
  • retro-daily

Business Sales

  • Drevon - Mac desktop workspace for GTM engineers. Run parallel AI agents powered by Claude Code, Codex, or Copilot to build target lists, score accounts, and pull prospect intel.
  • b2b-project-shipper
  • customer-success-manager
  • enterprise-onboarding-specialist
  • finance-tracker
  • mortgage - Mortgage refinance plugin by LendTrain — real-time institutional pricing, compliance, and FHA/VA loan detection via MCP. No API key required.
  • pricing-packaging-specialist
  • product-sales-specialist
  • support-responder
  • stratarts
  • technical-sales-engineer
  • founder-toolkit - 4 skills for startup founders: investor update writer (YC/a16z format), pitch deck reviewer (10-dimension VC scoring), SaaS metrics dashboard with benchmarks, and outreach lead scorer (ICP-first, tested on 300+ companies). Built by Mu Chen, CEO of Canopy Cloud ($120B+ AUM).
  • overloop-cli
  • signals-cli

Code Quality Testing

  • autoresearch
  • api-tester
  • bug-detective
  • claudex
  • code-review
  • code-review-assistant
  • code-reviewer
  • database-performance-optimizer
  • debug-session
  • debugger
  • double-check
  • optimize
  • performance-benchmarker
  • refractor
  • sdlc-wizard - SDLC enforcement plugin with hooks for TDD gates, planning workflow, confidence levels, and cross-model review. Installs via npx agentic-sdlc-wizard init.
  • test-file
  • test-results-analyzer
  • test-writer-fixer
  • unit-test-generator
  • vibe-guard - Always-on security guardrails for AI-generated code
  • think-first - Plan-before-code behavior modifier for agents
  • sonmat — Verification discipline plugin with six reactive axes (guard, inspect, witness, punch, devil’s advocate, scribe) for AI-human collaboration.
  • slm-mcp-hub - External: Universal MCP gateway federating 430+ tools from 37 servers through one endpoint. 75% RAM reduction.
  • tailtest - Automatically generates and runs tests for every file Claude Code creates or modifies. PostToolUse hook detects changes, generates test scenarios, runs them, and surfaces failures. 8 languages (Python, TypeScript, JavaScript, Go, Ruby, PHP, Java, Rust). Zero config, zero commands.
  • adamsreview
  • slicewise
  • bullpen

Communication & Integrations

  • whatsapp-claude-plugin — WhatsApp channel plugin for Claude Code. Connects as a linked device via Baileys v7 with bidirectional messaging, full media support, voice transcription, permission relay, and access control.

Data Analytics

  • alphai
  • analytics-reporter
  • crypto-trading-desk
  • data-scientist
  • dna-claude-analysis
  • experiment-tracker
  • feedback-synthesizer
  • newsmcp — Real-time world news for AI agents — events clustered from hundreds of sources, classified by 12 topics and 30+ geographic regions, ranked by importance. Free, no API key required.
  • trend-researcher
  • wellnizz
  • xpoz-brand-intelligence - Brand sentiment, competitive analysis, and influencer discovery using real-time Twitter/X data
  • x-twitter-scraper
  • claude-ecom
  • XVARY Stock Research — Claude Code skill for public SEC EDGAR + market data: /analyze, /score, /compare. MIT.
  • flex-claudecode
  • ai-dev-jobs-search
  • nhs-mcp-search
  • session-tax

Audio & Media

  • bibigpt-skill - AI-powered video, audio & podcast summarization

Design UX

  • brand-guardian
  • joker
  • journey-mapper
  • mobile-ux-optimizer
  • nano-banana - Google Gemini image generation plugin. Text-to-image, text-guided image editing, style transfer, 4K output, search grounding, and multi-reference composition via /genimage. Uses gemini-2.5-flash-image (fast) and gemini-3-pro-image-preview (4K/search).
  • onomastophes
  • ui-designer
  • ux
  • ux-researcher
  • visual-storyteller
  • whimsy-injector
  • claude-bionify

Development Engineering

  • claude-sounds
  • ai-engineer
  • claw-army/claude-node - Python subprocess bridge for Claude Code CLI, giving Python code direct access to Claude Code native capabilities via stream-json.
  • api-integration-specialist
  • backend-architect
  • blueprint
  • code-architect
  • context-memory
  • desktop-app-dev
  • enterprise-integrator-architect
  • flutter-mobile-app-dev
  • frontend-developer
  • ios-app-dev-skills
  • mobile-app-builder
  • html-report
  • project-curator
  • python-expert
  • rapid-prototyper
  • react-native-dev
  • silicon-friendly
  • vision-specialist
  • web-dev
  • tailwind-best-practices - Tailwind CSS v4 patterns — stop agents from generating v3 code
  • typescript-best-practices - Modern TypeScript — strict mode, discriminated unions, satisfies
  • python-best-practices - Modern Python 3.12+ — type hints, async, Pydantic v2
  • sveltekit-best-practices - SvelteKit — Svelte 5 runes, load functions, form actions
  • drizzle-best-practices - Drizzle ORM — schema design, relations, type-safe queries
  • fastapi-best-practices - FastAPI — async, dependency injection, Pydantic v2
  • shadcn-best-practices - shadcn/ui — component customization, forms, theming
  • expo-best-practices - Expo/React Native — navigation, EAS Build, platform splits
  • flutter-best-practices - Flutter/Dart 3 — GoRouter, Riverpod, sealed classes
  • go-best-practices - Modern Go — error wrapping, slog, context, goroutine lifecycle
  • image-link
  • agentkits-memory
  • lightcms
  • deapi-media
  • reporecall - Local codebase memory for Claude Code. AST indexing (22 languages), call graphs, hybrid search. Injects context via hooks (~5ms). Also an MCP server.
  • dotforge - Configuration factory for Claude Code — 17 skills, 7 agents, 15 stacks, audit scoring (0-10), practices pipeline.
  • slack-message-formatter
  • cc-hud — Compact single-line statusline: model name, context usage bar, active subagents, rate limits. Zero dependencies.
  • vara-skills
  • superlocalmemory - External: Persistent memory + learning MCP server for Claude Code. Zero-cloud, 6-channel retrieval, 16K+ monthly downloads. Backed by 3 arXiv papers.
  • skill-auto-installer
  • tldr
  • Imagine - Brings Google Gemini into Claude Code & Codex: delegate reasoning, independent code review, deep research, and automatic prompt-engineering. Runs on your Google AI Pro subscription, not your agent’s tokens.

Documentation

  • analyze-codebase
  • changelog-generator
  • codebase-documenter
  • context7-docs-fetcher
  • dev-report
  • documentation-generator
  • generate-api-docs
  • openapi-expert
  • update-claudemd
  • readme-best-practices - Stop agents from writing dry API-doc READMEs
  • prompt-compression - Compress docs/prompts into minimal tokens for AGENTS.md
  • conversation-memory - Give agents memory across sessions
  • ai-humanizer - Remove AI-detectable patterns from generated content
  • llm-wiki - Build a compounding, LLM-curated personal knowledge base in any project (Karpathy’s LLM Wiki pattern). Skill + 5 /wiki:* commands with sharded indexes, BM25 search, and a structural lint script.
  • swarmvault
  • verifying-phd-citations
  • thermal-fluid-research-workflow
  • trigger-tree

Git Workflow

  • analyze-issue
  • bug-fix
  • commit
  • create-pr
  • create-pull-request
  • create-worktrees
  • fix-github-issue
  • fix-issue
  • fix-pr
  • github-issue-fix
  • husky
  • pr-issue-resolve
  • pr-review
  • update-branch-name

Marketing Growth

  • app-store-optimizer
  • claude-rank - SEO/GEO/AEO audit with 170+ rules, auto-fix for robots.txt/sitemap.xml/llms.txt/JSON-LD
  • content-creator
  • growth-hacker
  • Hermes Tweet - Native Hermes Agent plugin for X/Twitter research, monitoring, drafting, follower exports, and approval-gated actions.
  • instagram-curator
  • poppify — photo-led short-form vertical reels (Instagram/TikTok/YouTube Shorts/Facebook) via MCP, $0.06/render, 50 free seeds, no subscription
  • reddit-community-builder
  • taisly-agent-kit
  • tiktok-strategist
  • toprank
  • twitter-engager
  • wondelai-skills
  • agentkits-marketing
  • claude-code-marketing-skills - AI marketing skills: SEO Audit, Landing Page Review, Competitor Analysis, Ad Copy Writer, Lead Qualification. 5 free + premium via Cogny MCP
  • unslop
  • claude-persona
  • Pulse - Claude Code sub-agent that scores your X (Twitter) drafts against the published twitter/the-algorithm heavy-ranker weights (+75 reply_engaged_by_author, -74 negative_feedback_v2, link tax, etc.) before you post. Outputs a 0-100 score, 5-axis breakdown, ship/rewrite verdict, and 3-5 concrete rewrites cited to which axis lifts. Drops into ~/.claude/agents/pulse.md. Open-source sample: github.com/septimlabs-code/septim-agents-pack-sample. $49 founding rate thru Mon 2026-05-26 ($79 after). septimlabs.com/pulse
  • social-media-publisher
  • Diffmode Growth Tactics — free plugin + marketplace that runs a growth-strategy pipeline: a competitor read, a buyer jobs-to-be-done map, and 7–9 unconventional acquisition tactics for bootstrapped founders. Claude Code + Codex, Apache-2.0.
  • notfair
  • social-vision
  • x-skills
  • yaohe

Project & Product Management

  • bleu
  • discuss
  • explore
  • fractal - Recursive project management plugin. Decomposes any goal into verifiable predicates, works on the riskiest unknown first. Features /fractal:run (idempotent state machine), /fractal:init, /fractal:patch, dry run mode, and incremental decomposition with re-evaluation.
  • plan
  • planning-prd-agent
  • plannotator - Interactive plan review UI with visual annotation (install via /plugin marketplace add backnotprop/plannotator)
  • prd-specialist
  • product-manager-skills - Senior PM agent with 30+ frameworks, 32 SaaS metrics with exact formulas, 12 templates, and anti-pattern detection. Pure Markdown, MIT-0 license.
  • project-shipper
  • sprint-prioritizer
  • studio-producer
  • tool-evaluator
  • tycana
  • workflow-optimizer
  • prd-generator — Self-contained PRD generator from conversation context
  • agent-decision-record
  • product-org-os
  • governor

Lifestyle & Entertainment

  • ai-divination-skills

Security, Compliance, & Legal

  • ai-ethics-governance-specialist
  • audit
  • compliance-automation-specialist
  • data-privacy-engineer
  • enterprise-security-reviewer
  • security-sweep - Comprehensive security scanner covering OWASP Top 10 (2025), Mobile Top 10 (2024), and LLM Top 10 (2025). Scans for hardcoded secrets, injection flaws, auth issues, misconfigurations, AI-specific vulnerabilities, mobile security, and data exposure.
  • legal-advisor
  • legal-compliance-checker
  • openclaw-security - 11-tool security suite for agent workspaces: integrity scanning, secret detection, permission auditing, network DLP, supply-chain checks, injection defense, credential lifecycle, compliance enforcement, audit trails, cryptographic signing, and incident response.
  • privacy-mask - Detect and redact PII, API keys, and secrets in screenshots via OCR + regex before images reach Claude. 47 detection rules, Tesseract/RapidOCR support, UserPromptSubmit hook integration. 100% offline.
  • skillfortify - External: Formal verification scanner for AI agent skills (including Claude Code skills + MCP servers). 100% precision on 540-skill benchmark. ASBOM generation. arXiv:2603.00195.
  • shellgate
  • huiyu-safe-ai - Lightweight AI security guard for install/download commands. Blocks 68+ malicious packages, verifies identity, scans code in <1s. Built from a real supply chain attack.
  • web-security-guard
  • supply-chain-gate
  • agent-safety-preflight

MCP Servers

  • AccInt — Local-first Work Model MCP server and Claude Code/Codex/OpenCode plugin. Tools: acc_retrieve, acc_act; official registry io.github.maxbaluev/accint.
  • lazymac/mcp — Unified MCP server exposing 42+ developer tools (qr, ip-geo, ai-cost, llm-router, k-privacy, korean-nlp) backed by Cloudflare Workers. npx -y @lazymac/mcp
  • lazymac/k-mcp — Korean wedge MCP — PIPA compliance, KRW + BOK rates, 사업자등록번호 lookup, address geocoding, NLP. npx -y @lazymac/k-mcp
  • Lobex - Agent-to-agent marketplace. Remote MCP https://lobex.app/mcp + stdio client. Site: https://lobex.app
  • mcp-interactive-terminal — Real interactive terminal sessions for REPLs, SSH, database clients, and any interactive CLI — with clean text output, smart completion detection, and multi-layer security. Install: claude mcp add terminal -- npx -y mcp-interactive-terminal
  • SkillFlow - AI Skills Marketplace discovery server. Search, compare, and get installation instructions for curated MCP servers and AI agent tools.
  • WhichModel — AI model pricing & recommendation MCP server for Claude Code. Helps choose the right model for every task at the best price. Cross-verified data updated every 4 hours. MCP endpoint: https://whichmodel.dev/mcp
  • Synder Importer MCP — Official Synder plugin. Import CSV/XLSX accounting data into QuickBooks Online or Xero via the Synder Importer API. 19 MCP tools covering imports, field mapping rules, post-import rules, and entity discovery. Bundles the gl-importer agent skill.
  • AgentIQ / MoltAd — Publisher MCP: list placements, deliver_ad, earn credits. https://moltad.net/publishers · https://moltad.net/mcp

Thinking & Knowledge Management

  • thinking-tree

Companion Apps & Tools

  • Onepilot — iOS app to SSH into remote servers and run Claude Code from your phone

Knowledge Management

  • bedrock

External Marketplaces

Community-maintained plugin marketplaces you can add to access additional plugins.

Marketplace Description Install
ykdojo/claude-code-tips dx plugin: GitHub Actions analysis (/gha), conversation cloning (/clone, /half-clone), context handoffs (/handoff), Reddit fetching (/reddit-fetch) claude plugin marketplace add ykdojo/claude-code-tips then claude plugin install dx@ykdojo

Skills & Frameworks

  • aurakit — All-in-one Claude Code skill: 46 modes, 23 sub-agents, 6-layer OWASP security, 10 lifecycle hooks, ~55% token savings. Cross-platform (Codex, Cursor, Manus, Windsurf). Install: npx @smorky85/aurakit

Resources

  • Claudebin (GitHub) - A minimalistic tool for publishing and sharing Claude coding sessions.

Tutorials

You can host and share your own curated plugin marketplace using a simple Git repository with a .claude-plugin/marketplace.json file. Users can install your marketplace via:

# /plugin marketplace add ccplugins/marketplace
/plugin marketplace add user-or-org/repo-name

Then browse and install plugins from within Claude Code’s /plugin menu.

Example:

/plugin
/plugin install analyze-codebase

Marketplaces

  • AgentStore - Open-source plugin marketplace with gasless USDC payments. Install via /plugin marketplace add techgangboss/agentstore. Publishers earn 80% of sales. Agent-first API for zero-auth publishing.
  • protonium — AI agents and productivity tools. Includes AxiomCore, a project & routine management agent: enforced structure (numbered folders, task IDs, wiki memory), plan → approve → execute workflow, agile or WBS project creation.

Contributing

Contributions are welcome! You can add your favorite plugins, share best practices, or submit your own marketplace.

评论

登录 后发表评论
暂无评论
GitRead 中文社区
深度解读 · 趋势洞察 · 价值精选

关于

关于我们联系我们编辑准则收录标准赞赏人员

发现

技术趋势新手训练营资源精选实战指南精选专题热门话题GitHub 文档

条款

隐私政策服务条款网站地图
26.9万收录项目
5万中文话题
155精选专题
发现技术价值,连接开源智慧
© 2026 GitRead · GitHub 中文社区 本站与 GitHub, Inc. 无隶属关系;GitHub 是 GitHub, Inc. 的商标。