{
  "markdown": "# ✨ Magic Powers\n\n[![npm version](https://img.shields.io/npm/v/magic-powers)](https://www.npmjs.com/package/magic-powers)\n[![License: MIT](https://img.shields.io/badge/license-MIT-blue)](LICENSE)\n[![Security Audit](https://img.shields.io/badge/security_audit-passing-brightgreen)](SECURITY.md)\n[![Tools](https://img.shields.io/badge/tools-9-green)](README.md#multi-tool-support)\n[![Agents](https://img.shields.io/badge/agents-16-purple)](README.md#16-built-in-agents)\n[![Commands](https://img.shields.io/badge/commands-15-blue)](README.md#15-slash-commands)\n[![Website](https://img.shields.io/badge/website-magic--powers.pmai.space-orange)](https://magic-powers.pmai.space)\n\n**[magic-powers.pmai.space](https://magic-powers.pmai.space)** — A Claude Code plugin with cost-optimized model routing, 43 built-in skills + 67 optional skills + 42 cloud division skills (232 total), 16 built-in agents + 33 optional agents, 15 slash commands, and optional hooks/MCP integration. Built for all professional roles — dev, product, data, marketing, sales, design, and more. Also works with Cursor, Copilot, Aider, Windsurf, Gemini CLI, Codex, Kiro, and OpenCode.\n\n## Why Magic Powers?\n\nClaude Code defaults to using the most expensive model for everything. Magic Powers routes tasks to the right model — **~75% cost reduction** with no quality loss on routine tasks.\n\nFor actual API-level savings, combine with [Claude Code Router or OpenRouter](docs/MODEL_ROUTING.md) for **80-95% cost reduction**.\n\n## Install\n\n```bash\n# Claude Code — marketplace (recommended)\n/plugin marketplace add kienbui1995/magic-powers\n/plugin install magic-powers@magic-powers\n\n# skills.sh (works with 43+ agents)\nnpx skills add kienbui1995/magic-powers\n\n# All other tools — npx (auto-detects your tool)\nnpx magic-powers@latest\n\n# Or with curl\ncurl -fsSL https://raw.githubusercontent.com/kienbui1995/magic-powers/main/scripts/get.sh | bash\n```\n\nThe installer auto-detects Cursor, Windsurf, Kiro, Codex, OpenCode, Gemini CLI, and Claude Code from your environment.\n\n### Codex\n\n```bash\ngit clone https://github.com/kienbui1995/magic-powers.git\ncd magic-powers && bash scripts/install.sh\n# Select option 7 — Codex\n```\n\n### Kiro\n\n```bash\ngit clone https://github.com/kienbui1995/magic-powers.git\ncd magic-powers && bash scripts/install.sh\n# Select option 8 — Kiro\n```\n\n## Built-in vs Optional\n\nMagic Powers uses a two-tier architecture:\n\n### Built-in (auto with plugin install)\n\nEverything works out of the box — no setup required:\n\n- **15 slash commands** — `/brainstorming`, `/plan`, `/review`, `/debug`, `/deploy`, etc.\n- \\*\\*16 built-in agents\\*\\* — architect, debugger, reviewer, ui-designer, etc.\n- **43 built-in skills** — brainstorming, TDD, debugging, security review, etc.\n- **SessionStart hook** — auto-loads skill routing on every session\n\n### Optional (via `/setup`)\n\nRun `/setup` to personalize for your project. Detects your stack, asks your role & priority, then offers:\n\n| Feature | Type | Description |\n|---------|------|-------------|\n| 🛡️ Safety guard | Hook | Block writes to `.env`, secrets, `node_modules` |\n| 🔍 Auto-lint | Hook | Run linter after every file edit |\n| 🧪 Auto-test | Hook | Run related tests after code changes |\n| 🎨 Stitch Design | MCP | Generate UI designs (needs `STITCH_API_KEY`) |\n| 📚 Context7 | MCP | Fetch latest library docs automatically |\n| 📋 Project conventions | Skill | Coding rules based on detected stack |\n| 🏗️ Stack-aware agents | Agent | Architect & debugger with stack context |\n| 📦 Optional skill packs | Skills | 67 role-specific skills across 11 categories |\n\nOptional features install to `.claude/` in your project — they don't affect the global plugin.\n\n### Optional Skills (67 skills, 11 categories)\n\nBrowse and install with `/install-skills`. Quick role-based install during `/setup`.\n\n| Category | Skills | Who it's for |\n|----------|--------|-------------|\n| 🎯 Product | 6 | Product managers, founders |\n| 📊 Data/ML | 9 | Data scientists, ML engineers |\n| ⚙️ Platform/SRE | 6 | SRE, platform, DevOps |\n| 🎨 Design/UX | 4 | Product designers, UX researchers |\n| 👥 Team Processes | 9 | Team leads, eng managers |\n| 📣 Marketing | 8 | Marketers, growth engineers |\n| 💼 Sales | 5 | Sales, BD, account executives |\n| 🧪 Testing | 5 | QA engineers, SDETs |\n| 🎮 Game Dev | 5 | Game designers, developers |\n| 🌐 Spatial Computing | 3 | XR/AR/VR, visionOS developers |\n| 🔬 Specialist | 7 | Legal, finance, devrel, solutions arch |\n\nSee [docs/OPTIONAL_SKILLS.md](docs/OPTIONAL_SKILLS.md) for the full catalog.\n\n### ☁️ Cloud Divisions (42 skills, 21 agents)\n\nProfessional certification-aligned agents for GCP, AWS, and Azure. `/install-skills` → Category 12.\n\n| Division | Agents | Certifications |\n|----------|--------|---------------|\n| GCP | 7 | Professional Data Engineer, Cloud Developer, Network, ML, DevOps, Security, Architect |\n| AWS | 7 | Data Engineer (DEA-C01), Developer, Networking, ML Engineer (MLA-C01), DevOps, Security, Solutions Architect |\n| Azure | 7 | Fabric Data Engineer (DP-700), AI Developer (AI-200), Network, AI Engineer, DevOps, Security (SC-500), Architect |\n\n```\n/install-skills → 12 → GCP / AWS / Azure\n@gcp-data-engineer  Design a BigQuery streaming pipeline\n@aws-solutions-architect  Review our multi-account architecture\n@azure-ai-engineer  Build a RAG system with Azure OpenAI\n```\n\n### 📊 Amplitude Division (26 skills, 5 agents)\n\nProduct analytics agents powered by the Amplitude MCP server. `/install-skills` → Category 13.\n**Requires:** [Amplitude MCP](https://github.com/amplitude/mcp-server-guide) configured.\n\n| Agent | Role |\n|-------|------|\n| `@amplitude-analyst` | Charts, dashboards, daily/weekly briefings |\n| `@amplitude-experimenter` | A/B tests, opportunity discovery, user journeys |\n| `@amplitude-engineer` | Analytics instrumentation, event taxonomy |\n| `@amplitude-ux-researcher` | Session replay, error diagnosis, UX audit |\n| `@amplitude-ai-monitor` | AI/LLM quality, topic analysis, session investigation |\n\nSkills ported from [amplitude/mcp-marketplace](https://github.com/amplitude/mcp-marketplace).\n\n## 15 Slash Commands\n\n| Command | Skill | Purpose |\n|---------|-------|---------|\n| `/setup` | — | Personalize plugin, install optional features (12 roles) |\n| `/install-skills` | — | Browse and install optional skills by category |\n| `/brainstorming` | brainstorming | Explore ideas → design → spec → approval |\n| `/plan` | writing-plans | Break feature into implementation steps |\n| `/review` | requesting-code-review | Quick code review for bugs & issues |\n| `/debug` | systematic-debugging | Reproduce → isolate → fix → verify |\n| `/security-scan` | security-review | Scan for vulnerabilities & auth issues |\n| `/db-review` | database-optimization | Schema, queries, indexes, migrations |\n| `/deploy` | ci-cd-pipeline | Deployment readiness checklist |\n| `/pr` | pr-workflow | Prepare changes for pull request |\n| `/refactor` | refactoring | Improve structure, reduce complexity |\n| `/tdd` | test-driven-development | Red → green → refactor cycle |\n\n## 16 Built-in Agents\n\n### 🔧 Engineering\n\n| Agent | Emoji | Model | Purpose |\n|-------|-------|-------|---------|\n| `architect` | 🏗️ | Opus | Brainstorming, system design, planning |\n| `debugger` | 🐛 | Sonnet | Systematic debugging with full tool access |\n| `database-optimizer` | 🗄️ | Sonnet | Schema review, query optimization, migrations |\n| `sre` | 🔧 | Sonnet | Infrastructure, deployment, reliability |\n| `git-workflow` | 🌿 | Sonnet | Branch strategy, commits, release workflows |\n\n### 🎨 Design\n\n| Agent | Emoji | Model | Purpose |\n|-------|-------|-------|---------|\n| `ui-designer` | 🎨 | Sonnet | Frontend design with Stitch SDK |\n\n### 👀 Review & Quality\n\n| Agent | Emoji | Model | Purpose |\n|-------|-------|-------|---------|\n| `reviewer` | 👀 | Haiku | Fast code review (read-only) |\n| `security-reviewer` | 🛡️ | Haiku | Security audit, vulnerability scanning |\n| `technical-writer` | 📝 | Haiku | Documentation, READMEs, ADRs, changelogs |\n\n### 📣 Product & Growth\n\n| Agent | Emoji | Model | Purpose |\n|-------|-------|-------|---------|\n| `product-strategist` | 📣 | Sonnet | Feature prioritization, PRDs, launch planning |\n| `copywriter` | ✍️ | Haiku | Landing pages, marketing copy, announcements |\n| `solo-ai-builder` | 🚀 | Sonnet | Validation, positioning, GTM, monetization, retention |\n\n### 🧪 Quality & Testing\n\n| Agent | Emoji | Model | Purpose |\n|-------|-------|-------|---------|\n| `qc-engineer` | 🧪 | Sonnet | Test design, automation, defect management, quality metrics |\n| `qa-engineer` | 🎯 | Sonnet | QA process design, test strategy, risk management, audits |\n\n### 🔷 Azure DevOps\n\n| Agent | Emoji | Model | Purpose |\n|-------|-------|-------|---------|\n| `ado-cicd-engineer` | ⚙️ | Sonnet | Azure DevOps pipelines, CI/CD, release management, containers |\n| `azure-devops-admin` | 🏢 | Sonnet | ADO org setup, access control, work tracking, artifacts |\n\n### Cost by Model\n\n| Model | Agents | Cost |\n|-------|--------|------|\n| Opus | 1 (architect) | $$$$$ |\n| Sonnet | 11 (debugger, db-optimizer, sre, git-workflow, ui-designer, product-strategist, solo-ai-builder, qc-engineer, qa-engineer, ado-cicd-engineer, azure-devops-admin) | $$ |\n| Haiku | 4 (reviewer, security-reviewer, technical-writer, copywriter) | $ |\n\n## 232 Skills\n\n**Core Workflow:** `using-magic-powers` · `brainstorming` · `writing-plans` · `executing-plans`\n\n**Planning:** `spec-driven-development` · `mvp-rapid-development`\n\n**Development:** `test-driven-development` · `systematic-debugging` · `verification-before-completion` · `refactoring` · `api-design` · `performance-optimization` · `environment-setup`\n\n**DevOps:** `ci-cd-pipeline` · `docker-containerization` · `incident-response` · `dependency-management`\n\n**Collaboration:** `requesting-code-review` · `receiving-code-review` · `subagent-driven-development` · `dispatching-parallel-agents` · `pr-workflow`\n\n**Git:** `using-git-worktrees` · `finishing-a-development-branch`\n\n**Security & Auth:** `security-review` · `authentication-patterns` · `accessibility-compliance`\n\n**Infrastructure:** `database-optimization` · `infrastructure-review` · `caching-strategy`\n\n**AI Engineering:** `prompt-engineering` · `rag-architecture` · `agentic-ai-patterns` · `llm-evaluation` · `ai-safety-guardrails` · `llm-observability`\n\n**Review & Docs:** `technical-writing`\n\n**Meta:** `writing-skills` · `open-source-project` · `product-strategy`\n\n**Unique:** `cost-aware-routing` · `design-with-stitch` · `design-with-pencil`\n\n## Multi-Tool Support\n\n| Tool | Format | Install |\n|------|--------|---------|\n| Claude Code | Plugin (native) | `/plugin marketplace add kienbui1995/magic-powers` |\n| Cursor | `.mdc` rules | `bash scripts/install.sh` → select Cursor |\n| GitHub Copilot | Agent `.md` files | `bash scripts/install.sh` → select Copilot |\n| Aider | `CONVENTIONS.md` | `bash scripts/install.sh` → select Aider |\n| Windsurf | `.windsurfrules` | `bash scripts/install.sh` → select Windsurf |\n| Gemini CLI | Skills | `bash scripts/install.sh` → select Gemini |\n| Codex | Skills + `AGENTS.md` | `bash scripts/install.sh` → select Codex |\n| Kiro | Steering files | `bash scripts/install.sh` → select Kiro |\n| OpenCode | `AGENTS.md` | `bash scripts/install.sh` → select OpenCode |\n\n## Stitch Design Integration\n\n```bash\n# Requires STITCH_API_KEY environment variable\nnode scripts/stitch.mjs generate <projectId> \"A dashboard with sidebar\"\nnode scripts/stitch.mjs get-html <projectId> <screenId>\n```\n\n## Examples\n\nSee [`examples/`](examples/) for real-world scenarios:\n- [Feature Development](examples/01-feature-development.md) — full lifecycle with cost breakdown\n- [Production Debugging](examples/02-production-debugging.md) — incident response\n- [Database Review](examples/03-database-review.md) — schema + security review\n- [UI Design with Stitch](examples/04-ui-design-with-stitch.md) — design-to-code\n\n## vs. Superpowers\n\n| Feature | Superpowers | Magic Powers |\n|---------|-------------|--------------|\n| Skills | 14 | 232 |\n| Agents | 1 | 16 built-in + 33 optional |\n| Commands | 0 | 15 |\n| Model routing | ❌ | ✅ Opus/Sonnet/Haiku |\n| Cost optimization | ❌ | ✅ ~75% reduction |\n| Optional hooks | ❌ | ✅ Safety, lint, test |\n| Design tools | ❌ | ✅ Google Stitch SDK |\n| Multi-tool | ❌ | ✅ 9 tools |\n| Built-in/Optional arch | ❌ | ✅ |\n\n## Changelog\n\nSee [CHANGELOG.md](CHANGELOG.md) for release history.\n\n## Contributing\n\nSee [CONTRIBUTING.md](CONTRIBUTING.md) for how to add agents and skills.\n\n## License\n\nMIT — see [LICENSE](LICENSE).\n\n## Credits\n\nInspired by [superpowers](https://github.com/obra/superpowers) by Jesse Vincent and [agency-agents](https://github.com/msitarzewski/agency-agents) by Mike Sitarzewski.\n",
  "bytes": 12727,
  "sha": "0c31cad87542a69e64a7e8dfb8faba198d5aad8ac61f743ae7859faae9f92f30",
  "repo_slug": "kienbui1995/magic-powers",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_kienbui1995_magic_powers_magic_powers_b1671b3f/readme"
}