God-Level Skill Suite
ArdurAI/god-skill-suite
52 skills battle-tested para IA código sem alucinações
Ver no GitHubConjunto de 52 skills de produção para Claude Code, Cursor, Windsurf e outras IDEs com IA. Cada skill reposiciona o modelo como um especialista em engenharia de software, eliminando alucinações e entregando código verificado. Cobre arquitetura, DevOps, segurança, frontend, backend e cross-domain mastery.
Instalação
npx convez-ai ArdurAI/god-skill-suite1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Aumentar qualidade de código gerado por Claude Code, Cursor e Windsurf
- Eliminar alucinações em APIs, sintaxe e decisões arquiteturais
- Padronizar expertise de engenharia em equipes de desenvolvimento
Veredito

Suite madura e prática que resolve um problema real: alucinações em assistentes de código. As 52 skills são rigorosamente testadas e seu design anti-hallucination (forçar explicitação de incerteza) é uma abordagem sólida. Recomendado para times que usam Claude Code ou Cursor em produção.
README
A tradução para o português deste README ainda está na fila.
🌌 God-Level Skill Suite
52 Battle-Tested, Production-Grade AI Skills. Covering every dimension of modern software engineering. Zero Hallucination.
Built natively for Claude Code, Cursor, Windsurf, Continue, Gemini CLI, and more.
⭐ If this makes your AI coding sessions dramatically better, please star the repo! It helps other engineers discover the suite and motivates us to keep adding world-class skills.
📢 Share it with your team, in your Discord, or on Twitter/X — the more engineers use these skills, the more battle-tested they become.
🤝 Want to contribute a skill? Read the Contribution Guide — even a single high-quality skill makes a huge impact.
🎯 The Philosophy Grid
| 🛡️ Zero Hallucination | ⚔️ Researcher-Warrior | 🌐 Cross-Domain Mastery |
|---|---|---|
| No guessing. Strict anti-hallucination protocols force the model to invoke explicit uncertainty clauses rather than invent syntax or fabricate API behaviour. | No passivity. Every skill adopts the persona of a relentless engineer who traces bugs to their fundamental root and deeply interrogates every architectural decision. | No role siloing. A frontend engineer must understand Kubernetes networking. A DevOps engineer must understand OAuth flows. Lines fade deliberately. |
⚡ How It Works
🚀 One-Minute Deployment
curl -fsSL https://raw.githubusercontent.com/gnanirahulnutakki/god-skill-suite/main/installer/install.sh | bash
irm https://raw.githubusercontent.com/gnanirahulnutakki/god-skill-suite/main/installer/install.ps1 | iex
Install safely from PyPI for global CLI access across your system.
# With uv (fastest)
uv tool install god-skill-suite && god-skills install
# With pipx
pipx install god-skill-suite && god-skills install
# Direct, no install required (run from cloned repo)
python installer/install.py
📚 The 52-Skill Registry
[!IMPORTANT] Always load
god-meta-conductorfirst. It is the baseline operating system — establishing Zero-Hallucination rules and scope-locking before any domain skill activates.
| Skill | Description |
|---|---|
god-meta-conductor | Master orchestrator — anti-hallucination, cross-domain thinking, researcher-warrior identity |
god-dev-core | End-to-end software engineering: DSA, OOP, SOLID/DRY/YAGNI, design patterns, clean code |
god-dev-research | Research methodology: finding papers, novelty checking, arXiv/ACM/IEEE, GitHub research |
god-dev-builder | Building production systems: architecture decisions, scaffolding, iterative delivery |
god-dev-codebase | Working with existing codebases: reading, debugging, refactoring, PR reviews |
god-research-review | Deep research review: literature synthesis, critical analysis, experimental design |
| Skill | Description |
|---|---|
god-security-core | STRIDE, OWASP Top 10, zero-trust, cryptography, SLSA, SBOM, Vault, container hardening |
god-security-cloud | Cloud attack kill chain, privilege escalation, GuardDuty/Defender/SCC, CIEM |
god-auth-protocols | OAuth 2.0 PKCE, OIDC, SAML 2.0, JWT, mTLS, SPIFFE/SPIRE, WebAuthn/passkeys |
god-iam-aws | AWS IAM: policies, roles, SCPs, permission boundaries, IRSA, IAM Access Analyzer |
god-iam-azure | Azure Entra ID: RBAC, Managed Identities, PIM, Conditional Access |
god-iam-gcp | GCP IAM: resource hierarchy, Workload Identity, Organization Policy, IAM Deny |
| Skill | Description |
|---|---|
god-devops-core | CI/CD pipelines, GitOps, Docker, Helm, deployment strategies |
god-devops-kubernetes | Kubernetes deep dive: pods/services/ingress/RBAC/HPA/PDB/NetworkPolicy |
god-linux-mastery | Linux internals, kernel/process/memory, eBPF/bpftrace, cgroups v2, systemd |
god-containers-advanced | OCI spec, OverlayFS, BuildKit, multi-arch, rootless containers, gVisor/Kata |
god-infra-as-code | Terraform (advanced HCL, state, modules), Pulumi, AWS CDK, Crossplane |
| Skill | Description |
|---|---|
god-web3-blockchain | Solidity, EVM internals, Rust (Solana), smart contract security, DeFi, gas optimization |
god-database-mastery | PostgreSQL internals, EXPLAIN ANALYZE, indexing, sharding, replication, MongoDB, Redis |
god-data-engineering | Spark, Flink, Airflow/Prefect, dbt, data modeling, lakehouse, streaming pipelines |
god-data-cleaning | Data preprocessing, feature engineering, outlier detection, imputation, pipeline design |
god-message-streaming | Kafka internals (ISR, EOS, Streams), RabbitMQ, SQS/SNS, NATS JetStream, Pulsar |
| Skill | Description |
|---|---|
god-frontend-mastery | React 18+, Next.js App Router, TypeScript, state management, Core Web Vitals, WCAG 2.2 |
god-ui-ux-design | Interface design, HSL color theory, 8pt grids, accessibility WCAG 2.2, Spring physics |
god-mobile-awareness | iOS/SwiftUI, Android/Compose, React Native, Flutter, PWA, push notifications |
| Skill | Description |
|---|---|
god-ai-prompting | System prompts, chain-of-thought, few-shot, constitutional AI, evaluation |
god-ai-architect | RAG, agents, tool use, multimodal, inference optimization, responsible AI |
god-llm-sdk | Claude SDK, OpenAI SDK, LangChain, LlamaIndex, streaming, function calling |
god-mlops-core | ML pipelines, model versioning (MLflow/W&B), serving, drift detection, A/B testing |
god-mlops-llm | LLM fine-tuning (LoRA/QLoRA), RLHF, evaluation (RAGAS/DeepEval), LLMOps |
god-ml-data-training | Data labeling, training loops, loss functions, distributed training, checkpointing |
| Skill | Description |
|---|---|
god-serverless-architecture | AWS Lambda, DynamoDB single-table, API Gateway, Step Functions, idempotency |
god-systems-design | CAP/PACELC, consensus (Raft/Paxos), consistent hashing, CQRS, Event Sourcing, Saga |
god-architecture-patterns | Microservices, DDD, hexagonal/clean/onion, strangler fig, event-driven, ADRs |
god-api-design | REST (Richardson Maturity), GraphQL (N+1, DataLoader), gRPC, API versioning |
| Skill | Description |
|---|---|
god-observability | OpenTelemetry, Prometheus, Grafana, distributed tracing (Jaeger/Zipkin), alerting |
god-sre-reliability | SLOs/SLIs/error budgets, toil reduction, capacity planning, incident response, GameDay |
god-performance-engineering | CPU/memory profiling, flame graphs, JVM tuning, Go pprof, k6/Gatling |
god-platform-engineering | IDP design, Backstage, golden paths, developer portals, paved roads |
| Skill | Description |
|---|---|
god-search-engineering | Elasticsearch/OpenSearch internals, BM25 tuning, analyzer pipelines, relevance scoring |
god-vector-databases | Pinecone, Qdrant, Weaviate, pgvector, Chroma, FAISS — ANN algorithms, HNSW, hybrid search |
god-edge-computing | Cloudflare Workers (KV/R2/Durable Objects), Lambda@Edge, Fastly, Vercel Edge |
god-networking | TCP internals (BBR/CUBIC), DNS, TLS 1.3, VPC design, BGP, eBPF, troubleshooting |
| Skill | Description |
|---|---|
god-compliance-governance | SOC 2 Type II, HIPAA, PCI-DSS v4.0, GDPR, ISO 27001:2022, NIST 800-53 |
god-cost-engineering | FinOps, cloud cost optimization, reserved instances, rightsizing, cost attribution |
god-project-management | Agile/Scrum, Kanban, OKRs, RICE scoring, DORA metrics, blameless postmortems |
god-testing-mastery | Unit/integration/E2E/contract/load/chaos/fuzz/mutation testing — full test pyramid |
god-git-workflow | Git internals, rebasing, bisect, reflog, monorepo, hooks, conventional commits |
god-devex-tooling | VS Code/JetBrains/Neovim mastery, tmux, fzf, ripgrep, debuggers, linters |
god-tech-support | Full-stack troubleshooting: reading stack traces, log analysis, network debugging |
god-tech-support | Kubernetes cluster triage, network debugging, systematic root-cause analysis |
🧪 Validated Ecosystem
Every skill ships with a Prompt Library (prompts/examples.md) and an adversarial test suite (tests.json) that runs against local Ollama models or frontier APIs to empirically prove Zero-Hallucination compliance.
# Run the full 52-skill test suite locally (requires Ollama)
python scripts/evaluate_skills.py --provider ollama --model qwen2.5:7b --output docs/TEST_RESULTS.md
→ See docs/TESTING.md for the full framework guide and docs/TEST_RESULTS.md for latest results.
📄 License
MIT License — Designed for builders who refuse to guess.