Show HN: Claude Skills Starter Kit – 198 modular skills for AI agents

Yevhen Shaforostov released the Claude Skills Starter Kit on GitHub, offering 20 modular skill files that turn Claude Code, Antigravity, and AI agents into senior engineers, architects, and product managers. The full suite of 198 production-ready skills is available for $49 on Gumroad, covering multi-agent crews, LangGraph, pgvector hybrid search, and safe tool governance, used daily at Adsy.com.

Drop-in domain skill files that turn Claude Code,Antigravity, andAI Agentsinto a Senior AI Engineer, Architect, and Product Manager. This repo contains 20 starter skills . If you want the complete battle-tested suite of 198 production-ready skills Multi-Agent Crews, LangGraph, pgvector Hybrid Search, GEO/AEO AI Search, OCR Document Parsers, and Safe Tool Governance used daily in production at Adsy.com https://adsy.com : 👉 Get the Full 198-Skill Pack on Gumroad $49 https://shaforostov5.gumroad.com/l/njzfuo | Skill | Category | Description | |---|---|---| agent-introspection-debugging | 🤖 AI Agents | Structured self-debugging workflow & synthetic failure recovery | langchain-agent-patterns | 🤖 AI Agents | Production LangChain LCEL, @tool schema generation & ReAct loops | multi-agent-crew-patterns | 🤖 AI Agents | Hierarchical multi-agent team topologies, DAG tasks & role separation | openai-agents-sdk | 🤖 AI Agents | Multi-agent handoffs, guardrails & Swarm patterns | signal-based-ai-agent | 🤖 AI Agents | Event-driven LangGraph architectures + Redis deduplication | pgvector-hybrid-search | 🔍 RAG & Search | Native vector embeddings + full-text TSVECTOR & Reciprocal Rank Fusion | enterprise-rag-knowledge-systems | 🔍 RAG & Search | Real-time COGS metering & per-tenant token analytics | geo-aeo-agent-optimization | 🌐 AI Search GEO | Answer-first formatting for Perplexity, ChatGPT Search & Gemini Overviews | airops-programmatic-content-engine | 🌐 AI Search GEO | Scalable content engine with schema.org JSON-LD grounding | tdd-workflow | ⚙️ Engineering | Test-driven development with red-green loops | code-review-and-quality | ⚙️ Engineering | Multi-axis automated code quality reviews | security-and-hardening | 🔒 Security | Pre-flight security checklists, OWASP guardrails & injection defense | security-review | 🔒 Security | API authorization, secrets protection & auth verification | api-design | 📐 Architecture | RESTful & GraphQL design standards, error codes & rate limiting | eval-harness | 🧪 Testing & Evals | Eval-driven development EDD framework for agent trajectories | mcp-server-patterns | 🔌 Integrations | Model Context Protocol MCP TypeScript/Python servers & Zod tools | prd-critic | 📋 Product | Automated PRD review against user outcomes, edge cases & testability | product-spec-generator | 📋 Product | Spec-driven Product Requirements Document PRD generator | prompt-engineer | 🧠 Prompts | Advanced prompt design, zero-shot CoT & system prompt caching | ui-ux-pro-max | 🎨 UI/UX Design | Design tokens, component specs & modern aesthetic guidelines | git clone https://github.com/yevhens-hue/claude-skills-starter-kit.git For Claude Code CLI: mkdir -p ~/.claude/skills cp -r claude-skills-starter-kit/skills/ ~/.claude/skills/ Or for Gemini / Antigravity IDE: mkdir -p ~/.gemini/config/skills cp -r claude-skills-starter-kit/skills/ ~/.gemini/config/skills/ When you ask Claude to write tests, design an API, optimize for AEO, or build a multi-agent graph, it will automatically load the matching SKILL.md into its context on demand. - Give a ⭐️ star on GitHub if you find this useful - Sponsor ongoing development: github.com/sponsors/yevhens-hue https://github.com/sponsors/yevhens-hue MIT © Yevhen Shaforostov https://github.com/yevhens-hue