Routing rules

158+ Claude Code specialist subagents organized into ten categories and installable by category plugin or individual Markdown agent.

Confidence: high

Complete coding-agent methodology using automatically triggered skills for brainstorming, worktrees, detailed plans, subagent development, TDD, review, and branch completion.

Confidence: high

Rust CLI proxy/hook that compresses noisy shell, git, test, build, cloud, and infrastructure command output before agents read it.

Confidence: high

Packs a repository into AI-friendly XML/Markdown/text with token counts, ignore rules, Secretlint filtering, remote-repo support, and Tree-sitter compression.

Confidence: high

Vectorless reasoning-based RAG that builds hierarchical document trees and lets LLMs reason over them instead of vector similarity/chunking.

Confidence: high

Self-hostable workspace where many coding-agent CLIs act as teammates: issues, squads, runtimes, skills, autopilots, logs, costs, and review gates.

Confidence: high

Official MCP steering-group reference servers: Everything, Fetch, Filesystem, Git, Memory, Sequential Thinking, and Time.

Confidence: high

Minimal end-to-end LLM training harness covering tokenizer, pretraining, SFT, RL, evaluation, inference, and scaling/speedrun scripts.

Confidence: high

Minimal autonomous LLM research loop: agent edits train.py, runs fixed five-minute experiments, checks val_bpb, and keeps/discards changes.

Confidence: high

Pre-release code-acting agent runtime executing generated JavaScript handlers inside hardware-isolated Hyperlight micro-VMs with explicit host capabilities.

Confidence: high

Unofficial /level-up skill that scans Claude Code setup signals, assigns one of ten maturity levels, and gives/builds a next-step roadmap.

Confidence: high

Hackathon FastAPI + Next.js prototype forecasting spending from calendar events with dashboard, savings challenges, mock banking, and optional Gemini coach.

Confidence: high

Minimal inference code for open-weight FLUX.2 image generation/editing, including Klein 4B/9B and larger dev variants.

Confidence: high

Local AI CRM framework built on a dedicated OpenClaw profile/gateway with web UI and CLI lifecycle.

Confidence: high

Engineering plugin whose core loop brainstorms, plans, works, simplifies, reviews, and stores solved-problem knowledge for future agent work.

Confidence: high

Large curated Agent Skills/plugin catalog plus a Composio-powered connect-apps plugin for authenticated actions across many external services.

Confidence: high

Official OpenAI Claude Code plugin for Codex read-only/adversarial reviews, rescue delegation, session transfer, and background job management.

Confidence: high

Enhanced fork of Claudia: Tauri desktop GUI for Claude Code projects/sessions, custom agents, usage analytics, MCP management, checkpoints, and CLAUDE.md editing.

Confidence: high

Small Karpathy-inspired guidelines package for thinking before coding, simplicity, surgical changes, and goal-driven verification across Claude/Cursor forms.

Confidence: high

macOS Electron overlay for Claude Code with multi-tab sessions, approval UI, history, local voice input, attachments, and skill marketplace.

Confidence: high

Vector-backed semantic code search that indexes repositories and retrieves relevant code via MCP for Claude Code, Codex, Gemini, Cursor and others.

Confidence: high

Community MCP server discovery repository; its root README was empty in the connector snapshot while translated indexes remained searchable.

Confidence: low

Curated Claude Code ecosystem index covering official guidance, runtimes/clients, status lines, design, security, orchestration, skills, memory, and observability.

Confidence: high

Community collection of Claude prompt examples across coding, writing, marketing, business, analysis, and education.

Confidence: medium

Persistent memory plugin that captures session observations, summarizes/stores them, supports hybrid search, and injects relevant context into later sessions.

Confidence: high

Claude Code/Cursor skills for Arcads image/video APIs, prompt libraries, static ad templates, animated pipelines, captions, and Meta-ad workflows.

Confidence: high

Public-beta autonomous engineering-team runtime coordinating planning, coding, review, QA, merge, verification, and replanning agents.

Confidence: high

Small, editable, composable engineering skills for requirements, specs/tickets, TDD, debugging, architecture, research, review, and implementation.

Confidence: high

Adaptive AI software-delivery method with specialized product, architecture, UX, development, and testing roles and right-sized workflows.

Confidence: high

Archived legacy GSD repository whose README directs users to open-gsd/gsd-core.

Confidence: high

Spec-driven development toolkit/Specify CLI with constitution → specification → plan → tasks → implementation plus clarify/analyze/checklist extensions.

Confidence: high

Converts books/document folders into structured Agent Skills with SKILL.md, chapters, glossary, patterns, and cheatsheet.

Confidence: high

Claude Code status-line HUD for project path, context health, tools, agents, todos, and usage from native session/transcript data.

Confidence: high

First-party Vercel CLI and deployment tooling monorepo for building/deploying web projects to the Vercel platform.

Confidence: high

Postgres development platform combining database, auth, REST/GraphQL APIs, realtime, storage, functions, vector/AI tooling, dashboard, and clients.

Confidence: high

Frontend design/quality Agent Skill plus commands, detector rules, PRODUCT.md/DESIGN.md artifacts, and live browser iteration.

Confidence: high

Python CLI/Agent Skill that installs, configures, and health-checks access paths for websites, search, social platforms, video, podcasts, RSS, and GitHub.

Confidence: high

Local-first desktop design studio connecting coding agents to prototypes, dashboards, decks, images, HyperFrames video, design systems, plugins, and exports.

Confidence: high

UI/UX design-intelligence skill with searchable styles, palettes, typography, charts, stack guidance, anti-patterns, and UX rules.

Confidence: high

Source-available visual workflow/AI-agent platform with a large connector ecosystem, templates, custom JavaScript/Python, self-hosting, and cloud.

Confidence: high

Portable frontend design Agent Skills spanning design taste, redesign, image-to-code, high-end/minimal/brutalist styles, and image/brand reference workflows.

Confidence: high

Web data API/open-source stack for search, scraping, crawling, browser interaction, and LLM-ready extraction.

Confidence: high

Local-first AI gateway routing coding agents across hundreds of providers/models with fallback, dashboard, MCP/A2A, and RTK/Caveman prompt compression.

Confidence: high

Python scraping framework with adaptive parsers, HTTP/browser/stealth fetchers, concurrent spiders, CLI, and MCP/agent integrations.

Confidence: high

Archon/Claude Code workflow that turns a topic or URL into a vertical AI-voiced HyperFrames short with HTML/GSAP templates and TTS.

Confidence: high

Official Anthropic Agent Skills examples and reusable skills for documents, design, development, communications, skill authoring, MCP building, and web testing.

Confidence: high

Anthropic's agentic coding CLI plus official plugins containing reusable commands, subagents, hooks, skills, and MCP integrations.

Confidence: high

Static-analysis CLI that extracts design systems from URLs, repositories, or local code and writes agent-ready design artifacts/skills.

Confidence: high

Teams-first multi-agent orchestration layer for Claude Code with guided setup, staged autopilot, native teams, provider advisers, planning, QA, memory, diagnostics, and many focused skills.

Confidence: high

Open-source agentic video-production system with 12 production pipelines, a large tool registry, provider selection, quality gates, project artifacts, and a live Backlot storyboard.

Confidence: high

Agent-integration platform and SDK monorepo providing toolkits, tool search, per-user authentication, sessions, triggers, and a sandboxed workbench across TypeScript and Python agent frameworks.

Confidence: high

Collection of 49 marketing Agent Skills spanning conversion, copy, SEO, paid acquisition, measurement, retention, growth engineering, strategy, monetization, sales, and RevOps.

Confidence: high

Portable writing skill that identifies and removes common signs of formulaic AI-generated prose while preserving meaning, facts, and quoted material.

Confidence: high

Local native code-intelligence MCP server that indexes repositories into a persistent structural knowledge graph and answers graph, architecture, impact, and source queries.

Confidence: high

Token-efficiency toolkit for coding agents: a concise-output skill plus a local proxy, recovery store, analyzer, browser, memory, and compression commands.

Confidence: high

Agent Skills for Obsidian workflows and open formats, including Obsidian Markdown, Bases, JSON Canvas, the Obsidian CLI, and Defuddle web-page cleanup.

Confidence: high