Anthropic plugins

Enforce a structured TDD workflow with parallel task execution, isolated git worktrees, and systematic code review. Guides root cause investigation before fixes, verifies tests and linters pass before claiming completion, and dispatches subagents for independent tasks like fixing multiple test failures simultaneously.

Equip AI coding agents with production engineering skills to handle full dev lifecycles: refine ideas to specs, implement via TDD slices, run tests/debug, perform multi-axis code reviews, optimize perf/security, automate CI/CD, and execute ship checklists.

Orchestrate 1,388 specialized AI skills in Claude Code to automate expert workflows for Azure SDK integrations, Odoo/Shopify configs, SEO audits, security pentests, full-stack scaffolding, agent building, and DevOps pipelines across Python, React, AWS, Kubernetes.

Build and deploy production-grade LLM applications with LangGraph for agent orchestration, advanced RAG pipelines leveraging vector and hybrid search, prompt engineering patterns, and automated evaluation. Covers embedding model selection, vector index optimization, and multi-agent architectures for document Q&A, chatbots, and semantic search over proprietary data.

Design, build, and deploy MCP servers for Claude by interrogating your use case to select deployment models like remote HTTP, MCPB, or local stdio, implementing tool patterns with auth, adding interactive UI widgets such as forms, pickers, and dashboards for inline chat rendering, and packaging into standalone Node or Python .mcpb bundles for local distribution without user toolchain.

Build and evaluate production-grade AI agents using LangGraph, RAG systems, MCP servers, and prompt engineering patterns—with behavioral testing and reliability monitoring.

Accelerate LLM application development with production-ready patterns for context window management, RAG pipelines, prompt caching, observability via Langfuse, and agent architectures.

Upgrade Claude AI integrations by migrating code, prompts, and API calls from Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, automatically updating model strings across Anthropic, AWS Bedrock, GCP Vertex AI, and Azure AI Studio platforms.

Run comprehensive SEO audits and optimizations: crawl up to 500 pages, analyze backlinks and Core Web Vitals, validate schema and hreflang, generate content clusters and sitemaps, and track regressions—all from Claude Code.

Orchestrate multi-LLM workflows across providers (Claude, Gemini, Codex, Copilot) for the full software lifecycle — research, architecture, implementation, code review, security audit, testing, and delivery — using structured Double Diamond phases and autonomous agents.

Enable Claude Code to autonomously curate its memory, promote proven patterns to permanent CLAUDE.md rules, and extract reusable skills from recurring solutions, improving code quality and automation over time.

Manage AI-driven development workflows with hierarchical task trees, dependency graphs, automated subtask expansion, PRD-to-task parsing, status tracking, and intelligent task orchestration via natural language commands.

Delegate full-stack development workflows to Claude via 213 specialized agents, commands, and skills: refactor code, generate tests/deployments/Dockerfiles/K8s manifests, audit security/performance, document APIs/onboarding, orchestrate Git ops, and apply patterns across JS/TS/Python/Rust/Go/Java stacks.

Direct AI coding agents to create or update promptfoo evaluation suites with configs, prompts, tests, deterministic assertions, and provider setups following best practices. Streamline LLM eval coverage, regression debugging, and new eval matrix generation in JavaScript or Python projects using OpenAI or Anthropic models.

Implement a structured, spec-driven development lifecycle with AI agents: discuss requirements, research, plan, execute code, verify, and manage milestones. Includes multi-model plan review, automated code review, test generation, UI auditing, project documentation, and session handoff for continuous AI-assisted development.

Run a complete AI-assisted coding workflow with self-correcting memory, persistent FTS5-indexed research wikis, auto-research loops, multi-LLM council deliberation, and 8 specialized agents that coordinate parallel sessions, enforce quality gates, audit context costs, and capture learnings across every session.

Build and run constrained browser agents that enforce a domain allowlist via CDP Fetch interception, letting you safely automate browsing tasks without granting raw shell or CDP access.

Run a law school legal clinic inside Claude — onboard students, conduct structured client intake, track deadlines with malpractice-aware warnings, generate court-ready documents and correspondence, manage a supervisor review queue, and produce end-of-semester transition memos. Integrates with Slack, Google Drive, CourtListener, and Lexis+ for case management and legal research.

Debug and fine-tune language models using the Tinker API: diagnose training pipeline issues, replicate research papers, run RL/SFT/DPO experiments, and monitor training logs—all from the command line.

Build and orchestrate advanced Claude Code agentic workflows by creating meta-prompts, subagents, hooks, MCP servers, slash commands, and skills; execute hierarchical plans, run autonomous coding loops, apply expert debugging and productivity frameworks like 5 Whys or Eisenhower Matrix, and audit components for compliance and quality.

Detect and rewrite AI-generated Korean text to sound human-written, using a multi-phase pipeline that scans for 40+ AI-typical patterns across 10 categories, preserves content, and validates semantic equivalence.

Orchestrate multi-agent AI systems with AI SDK v5 for task decomposition, handoffs, routing, and coordination across OpenAI, Anthropic, and Google providers. Use commands to initialize projects, generate specialized agents with custom prompts and tools, test workflows with metrics, and deploy orchestrator agents for complex task handling in TypeScript.

导入已有中文网络小说（半成品或完本）并自动解析为标准项目目录结构，支持长/短篇路由分流，打通从导入→分析→去AI化→封面设计→趋势研究→继续写作的完整创作管线，配合Agent协作与路由系统实现网文写作的逆向工程与正向产出闭环。

Monitor usage limits, reset times, and cost across multiple AI CLIs (Claude, Codex, Gemini, z.ai) from a unified terminal dashboard, with automatic recommendations for the CLI with most available capacity

Build production-grade LLM gateways with OpenRouter: route requests across 400+ models by task or criteria, chain fallbacks for reliability, cache responses to cut costs/latency, monitor usage/costs/latency, redact PII for compliance, and benchmark performance using Python OpenAI SDK wrappers.

Optimize LLM prompts for OpenAI and Anthropic by automatically detecting redundancy, simplifying instructions, and rewriting to reduce token usage, lower costs, and improve performance.

Accelerate Atomic Agents app development through a guided 7-phase workflow: delegate schema design, agent and tool creation, architecture planning, codebase analysis, and code review to specialized AI sub-agents for scalable multi-agent LLM systems.

Create and validate production-grade Claude Code skills per AgentSkills.io 2026 spec and 100-point rubric, plus Anthropic agent .md files matching 16-field 2026 standard. Audit existing skills/agents or build custom subagents for orchestrators and marketplace submission.

Master Cursor IDE AI workflows using 30 guided skills: install and authenticate, configure custom models and rules, optimize indexing and performance, automate Composer for multi-file refactoring and scaffolding, troubleshoot errors, manage teams with SSO, and audit compliance.

Rapidly build, debug, deploy, secure, monitor, and scale Lindy AI agents and multi-step workflows using 24 Claude Code skills that guide webhook integrations, CI/CD pipelines, error troubleshooting, cost/performance tuning, enterprise RBAC, and migrations from Zapier, n8n, or LangChain.

Execute deterministic, reproducible bioinformatics analyses locally — from genome annotation and pharmacogenomics to scRNA-seq and metagenomics — with built-in privacy and AI-driven skill orchestration.

Wrangle, profile, clean, transform, and analyze tabular data (CSV, TSV, Excel, JSONL, Parquet) using 51 qsv skill-based commands — including SQL queries, joins, validation, ontology inference, charting, performance acceleration, and reproducible logging.

Audit GitHub Actions workflows to detect security vulnerabilities in AI agent integrations like Claude Code Action, Gemini CLI, OpenAI Codex, and GitHub AI Inference. Identify prompt injection risks and unsafe input flows in CI/CD pipelines before deployment.

Integrate TwinMind AI to automate meeting transcription with speaker diarization, generate AI summaries and action items, sync tasks to Asana/Linear/Jira, handle webhooks/events, optimize costs/performance, configure security/RBAC, deploy to CI/CD/prod environments, and troubleshoot/migrate setups.

Delegate complex AI and data tasks to specialized agents that proactively build LLM applications with RAG and orchestration, design scalable ETL pipelines and warehouses, deploy MLOps workflows, optimize prompts, analyze datasets, manage context, and decompose goals into actionable hierarchies.

Query multiple AI coding agents (Gemini, OpenAI, Grok, Perplexity, and local Claude subagents) to get diverse perspectives on architecture decisions, debugging dead-ends, and technology choices, with automated synthesis of consensus and divergence.

Develop, deploy, and manage Databricks-based data pipelines, ML models, AI agents, and dashboards using code, SQL, and declarative tools for multi-environment workflows.

Log AI experiments interactively from the terminal by capturing tool usage, prompts, summaries, and ratings. Generate reports with stats, top tools, rating distributions, tags, and filters by tool, tag, days, or rating. Search experiments by query, tool, tag, rating, or date with sorting and stats display.

Analyze AI prompts for clarity, specificity, completeness, and issues with 1-10 scores and targeted fixes, then optimize by rewriting with structured best practices like sectioning, examples, chain-of-thought, and guardrails for superior LLM results.

Automates setup of MATLAB Agentic Toolkit: detects MATLAB installation, installs the MCP server, registers it with AI coding agents (e.g., Claude, ChatGPT), and verifies environment readiness.

Automates batch photo downloads from PhotoPlus and Alltuu cloud albums, while also providing a toolkit for configuring Claude Code agents, troubleshooting service deployments (CoPaw, OpenClaw, Hermes), generating structured documentation (PDF extraction, Mermaid diagrams, paper reports), and transcribing audio/video with MLX/Whisper.

Generate test fixtures that mock LLM responses, tool calls, errors, multi-turn loops, embeddings, and structured output across multiple AI providers for use with Copilot Kit's aimock library.

Refine rough prompts into precision-optimized versions for ChatGPT, Claude, and Gemini using the 4-D methodology. Receive the polished prompt, a summary of improvements made, and targeted tips to enhance future prompting.

Apply 39 research-backed thinking frameworks and mental models directly in your debugging, planning, estimating, and decision-making workflows—from systems archetypes and Bayesian reasoning to pre-mortems and first principles—so you systematically avoid cognitive biases, surface root causes, and choose higher-leverage actions without leaving the chat.

Build, deploy, and operate AI agents on AWS with Amazon Bedrock AgentCore: scaffold projects, connect tools via Gateway and MCP, orchestrate multi-agent systems, enforce Cedar policies, debug with traces and logs, evaluate quality, and harden for production with auth, IAM, rate limiting, and cold-start tuning.

Build and deploy AI agents to trade crypto, stocks, forex, and derivatives on Kraken via bash CLI: monitor markets, execute strategies like DCA, grid bots, basis trades, portfolio rebalancing; manage risks, staking, subaccounts with paper trading default and live opt-in safeguards. Integrates with Claude, Cursor, VSCode for stdio tool calls.

Switch Claude Code into a compressed Thai+English mode that cuts token usage by 60-75% in technical responses while preserving accuracy, and view real-time token usage and cost savings for the current session.

Execute end-to-end feature development via phased AI waves—DISCOVER products with JTBD interviews, DISCUSS requirements and UX journeys, DESIGN architectures with C4 diagrams, DEVOPS infrastructure with Terraform/K8s, DISTILL BDD tests, DELIVER TDD code—enforced by 23 agents, automated reviews, and quality gates for production-ready outputs.

Intercept and transform LLM API traffic with a mitmproxy-based server that supports OAuth authentication, sentinel key substitution, and model routing for OpenAI and Anthropic SDKs, and debug traffic flows through CLI tools or WireGuard capture.

Run a full digital marketing agency from Claude Code — plan campaigns, execute paid/SEO/email/social, audit performance, manage brands, and generate client deliverables across 158 skills and 25 agents.

Delegate image analysis, OCR text extraction, barcode/QR detection, and document processing to a vision expert agent using latest models like GPT-4V, Claude Vision, Mistral-OCR, Tesseract, and EasyOCR for efficient visual AI workflows.

Rapidly implement production-ready AI/ML features in apps: integrate LLMs with prompt engineering and response handling, build ML pipelines for recommendation systems, add computer vision for visual search, and enable intelligent automation using OpenAI, Anthropic, LangChain, Hugging Face, or Ollama.

Build, deploy, and manage Vercel projects including Next.js apps, AI features with the AI SDK, serverless functions, microfrontends, monorepos, and marketplace integrations. Automate CI/CD pipelines, configure environment variables, manage storage and authentication, and optimize performance and security

Build, configure, and deploy AI agents using the Motus framework through guided editor commands. Define ReAct agents with custom tools, workflows, persistent memory, and guardrails, then serve them locally or to cloud platforms.

Build, audit, and optimize Claude Code plugins using structured workflows for skill/hook authoring, TDD validation, quality scoring, security compliance, and performance analysis across plugin, project, and global scopes.

Rapidly implement production-ready AI/ML features in apps: integrate LLMs via prompt engineering and response handling, build ML pipelines for user behavior-based recommendations, add computer vision for photo-based product search, and deploy intelligent automations.

Refine rough prompts for AI platforms like ChatGPT, Claude, and Gemini into precision-optimized versions using the 4-D methodology, receiving the polished prompt, an improvements summary, and actionable tips.

Build, debug, and evaluate durable LLM workflows using the Output SDK — scaffold projects, manage encrypted credentials, orchestrate multi-step agents with Temporal, write and validate prompt files, and run cost analysis on workflow executions.

Refine product specifications iteratively through debates between multiple LLMs including Claude, OpenAI, Gemini, and Grok until consensus is reached. Activate interview mode for guided refinement, verify early agreements, and optionally integrate with Telegram for notifications.

Estimate per-turn token costs for .claude/ configurations and CLAUDE.md files, classifying components as always-loaded, path-scoped, or invoked-only, ranking top contributors, flagging overruns, and fetching exact counts via Anthropic API to optimize Claude context budgets.

Monitor Claude Code session costs and efficiency in real-time like htop, with breakdowns by git branch or PR, model pricing comparisons, spending analytics for custom periods, interactive web dashboard for trends and history, budget alerts, and CSV/JSON exports.

Automate desktop GUI workflows using Claude's Computer Use API: capture screenshots, control mouse and keyboard inputs, and run autonomous agent loops for tasks like form filling, visual app interactions, and GUI testing without CLI access.

Optimize Claude Code sessions by detecting/removing codebase bloat, dead code, and AI-generated hygiene issues; manage token budgets/context windows with MECW principles and subagent delegation; monitor CPU/GPU usage before intensive tasks; automate safe git-backed cleanups and audits.

Rapidly implement production-ready AI/ML features in apps, including LLM integrations with prompt engineering, ML pipelines for recommendations, computer vision for visual search, and intelligent automation, using a specialized agent.

Communicate in ultra-compressed mode to reduce token usage by ~75% while maintaining technical accuracy. Generate concise Conventional Commits, one-line PR reviews, and memory file compression for Claude Code sessions, with subagents for surgical code edits, codebase navigation, and diff auditing.

Orchestrate Claude Code into autonomous multi-session campaigns with fleet coordination, intent routing, and persistent state across context windows, plus skills for code review, testing, documentation, design manifest enforcement, and research.

Run adversarial debates between multiple LLMs to cross-validate code, architecture, reviews, security, research, and planning tasks, producing specialized outputs like code, reviews, or plans.

Assemble science-backed AI agent teams by decomposing complex missions into blueprints with topologies and roles, creating domain-specialized agents and skills, curating libraries for duplicates and quality, researching evidence-based briefs, and enforcing quality gates via structured reviews and verifications.

Generates business intelligence dashboards, infographics, AI images/videos, and automates posting to social media (WeChat, X, Xiaohongshu) with tools for document parsing, Obsidian integration, and browser automation.

Conduct comprehensive security audits and incident response across cloud, API, mobile, and AI systems with pre-built skills for compliance, threat modeling, and red teaming.

Apply Fable 5-level operational discipline—outcome-first communication, verification, tool parallelism, token economy—to Claude Sonnet sessions for lower-cost, higher-quality reasoning, code, and writing, plus wikilink-based prompt expansion via Python scripting.

Orchestrate full SDLC lifecycle phases from Inception through Transition using 58 AI agents and 170+ components to automate requirements, architecture evolution, testing orchestration, security gates, deployments, incident response, and project reporting via workflows, phase transitions, and quality checks.

Transform WPS Notes into a personal knowledge engine for long-form creative and academic writing. Automates memory retrieval, idea connection, insight generation, structured note-taking from any source (URLs, PDFs, images, audio transcripts), and multi-platform content formatting.

Runs operational-misalignment diagnostics on AI agents using iFixAi, guiding you from setup through fixture authoring to inspection with configurable judges from multiple model providers.

Run untrusted AI agent code in sandboxed environments with durable workflow orchestration, suspend/resume, and snapshot support — built for Python, TypeScript, and popular LLM frameworks.

Build RAG pipelines for document Q&A and chatbots by chunking large docs, generating embeddings, storing in vector DBs, and retrieving context to reduce hallucinations. Engineer and optimize LLM prompts using chain-of-thought, few-shot examples, constitutional AI, meta-prompting, and validation workflows.

Orchestrate persistent Claude Code agents across multiple communication channels (Telegram, WhatsApp, Discord, Gmail, X/Twitter), route messages, manage SWE tasks, and extend capabilities with custom skills, memory search, voice transcription, and PDF processing.

Enforces a structured development workflow through dual-perspective code review (Claude + Codex), automated test verification, security audits, and documentation sync. Guides bug fixes, feature implementation, and refactoring with state-machine gates and pre-commit safety checks.

A founder operating system that turns GitHub into a command center for running a business with AI agents — manages departments, weekly planning, retros, product roadmaps, competitive analysis, media production with HTML drafts, and Telegram bot operations through structured workflows and automated issue tracking.

Delegate advanced image analysis workflows to expert vision AI subagents that perform OCR with Tesseract/EasyOCR, barcode/QR detection, document processing, and optimization using cutting-edge models like GPT-4V, Claude Vision, and Mistral-OCR.

Automate full-stack development cycles: interview to generate executable PRDs, research codebase gaps, spawn agents to execute tasks/tests/docs in isolated git worktrees, iterate Greptile PR reviews until passing, and merge to main with health checks.

Generate synthetic datasets for LLM training using sdg_hub's composable blocks and YAML-defined flows, with support for 100+ LLM providers and custom scripts.

Persist and semantically search Claude AI conversation history across sessions and projects. Automatically detect continuity, predictively inject relevant context via hooks and tools, using a zero-dependency local embeddings server for instant retrieval without APIs or Docker.

Refine rough prompts into precision-crafted versions optimized for AI platforms like ChatGPT, Claude, and Gemini using the 4-D methodology. Instantly get the improved prompt, a summary of enhancements, and practical tips to boost AI response quality in your workflows.

Persist and recall Claude Code session context across conversations, learn from behavioral patterns, and enforce branch protection rules to prevent accidental edits to protected branches.

Exposes 200+ third-party API connectors (Salesforce, HubSpot, GitHub, Slack, Stripe, Jira, Notion, Intercom, Zendesk, and more) as strongly typed, documented tools so AI agents can call them via PydanticAI or the Anthropic/Claude SDK

Enforce a structured development workflow from ideation to code review: transform vague ideas into documented plans, verify code against specs, review git changes for SOLID violations and security risks, and guide test generation — all with Chain-of-Verification for accuracy.

Smith is a Spec-driven development harness for Claude Code with persistent memory, overnight batch builds, model routing, and agency operations tooling.

Index project documentation, codebases, and knowledge graphs for hybrid retrieval: BM25 keywords, semantic similarity, GraphRAG relationships, or fused multi-mode search. Retrieve cited chunks with scores to research dependencies, errors, and concepts in seconds using Ollama, OpenAI, or Anthropic.

Toggle Research Mode in Claude Code to enforce anti-hallucination safeguards: all responses require citations from local files or web searches, must ground claims in verified sources, and admit 'I don't know' for uncertainties. Persists across sessions until manually exited, with optional topic focus.

Consult Gemini 2.5 Pro, OpenAI Codex, and Claude for second opinions on debugging failures, architectural decisions, security validation, and code reviews. Assess coding problems to recommend and route to the best AI, gather repo context via git and files, invoke via CLI or subagents, and synthesize multi-AI analysis into actionable insights.

Build provider-agnostic, type-safe streaming LLM chats with tools, agent loops, and multimodal support directly in React and Next.js apps using hooks like useChat, compatible with OpenAI, Anthropic, Gemini, Ollama.

Orchestrate self-hosted AI agents via Station CLI to create agents, run tasks with 55+ MCP tools, manage environments, deploy teams, configure providers and backends through browser UI or CLI, and monitor executions with OpenTelemetry telemetry.

Polish AI-generated Korean text to sound more natural and human-like. Detects 40+ AI writing patterns across multiple categories, rewrites them with surgical precision, and validates that the original meaning and facts are preserved. Supports fast and strict modes, rollback, and versioned output.

Regenerate AIWG context files for Claude, Copilot, Cursor, Warp; scaffold agents, commands, skills, frameworks; manage workspaces with pruning, realignment, health checks; enforce @-mention traceability, validation, and linting; automate git commits, Docker/K8s deploys in agentic SDLC workflows.

Generate, modify, and style schema-driven React components from natural language prompts using Thesys AI. Outputs type-safe TypeScript code ready for Vite, Next.js, or Cloudflare Workers deployment.