By GadaaLabs
Enforces structured engineering workflows across the full development lifecycle: TDD discipline, root-cause debugging, multi-agent code reviews, git worktree isolation, and complexity-based model routing. Includes ML/AI/embedded domain patterns, UI/UX design assistance, and token cost monitoring.
Use when receiving code review feedback, before implementing any suggestion — requires technical verification and reasoned evaluation, not performative agreement or blind implementation
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
Use when starting any conversation - establishes how to find and use skills, requiring Skill tool invocation before ANY response including clarifying questions
Use when you have a spec or requirements for a multi-step task, before touching code
Store and retrieve patterns from past work using semantic search; adds self-learning capability to Superpowers
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
The complete, production-grade Claude Code skill system — packaged, optimized, and update-proof.
Built on obra/superpowers by Jesse Vincent. Extended with 10 new skills, specific engineering improvements directed by GadaaLabs, and an override infrastructure that makes your skills survive plugin updates automatically.
curl -fsSL https://raw.githubusercontent.com/GadaaLabs/claude-code-on-steroids/main/install.sh | bash
Requires: Claude Code CLI · Node.js 20+
Derived from Andrej Karpathy's observations on LLM coding mistakes. A lightweight addition to your
~/.claude/CLAUDE.mdthat works alongside the skill stack — no invocation required.
Why this complements the skill stack: Skills need to be invoked. A CLAUDE.md is ambient — it fires on every session automatically. These two principles fill the gap when a skill wasn't invoked:
| Principle | What it enforces | Covered by a skill? |
|---|---|---|
| Think Before Coding | Surface assumptions · present interpretations · ask before picking | Partial — architect covers design intent, not assumption surfacing |
| Goal-Driven Execution | Verifiable success criteria per step before writing code | Partial — forge + sentinel go deep, but only when invoked |
Add to your global CLAUDE.md in one command:
curl -fsSL https://raw.githubusercontent.com/GadaaLabs/claude-code-on-steroids/main/examples/karpathy-guidelines.md >> ~/.claude/CLAUDE.md
Full annotated example → examples/karpathy-guidelines.md
| obra/superpowers | Claude Code on Steroids | |
|---|---|---|
| Total skills | 14 | 24 |
| New skills not in obra | — | +10 (oracle, chronicle, vector, horizon, legion, pathfinder, gradient, nexus, ironcore, prism) |
| Domain expertise (ML/AI/EE/UI) | ✗ | ✓ |
| Intelligence layer (classifier, memory, router) | ✗ | ✓ |
| Multi-agent swarm templates | Concepts only | 4 topology prompt templates |
| API pre-verification in TDD | ✗ | ✓ (forge + blueprint) |
| Skill chain recipes | ✗ | 6 end-to-end chains |
| Domain trigger system | ✗ | 13 scenario → skill mappings |
| Override protection | ✗ | ✓ 100% update survivability |
| One-command install | ✗ | ✓ |
| Token usage analytics | ✗ | ✓ /tokenburn |
Run one command. The installer handles everything — installs the superpowers plugin, copies all 24 skills, builds the tokenburn CLI, and wires up the SessionStart hook:
curl -fsSL https://raw.githubusercontent.com/GadaaLabs/claude-code-on-steroids/main/install.sh | bash
Then open any project:
cd your-project && claude
/oracle # start here for every non-trivial task
/tokenburn # see where your tokens are going
Requirements: Claude Code CLI (npm install -g @anthropic-ai/claude-code) and Node.js 20+.
No need to uninstall anything. The installer detects your existing superpowers installation and installs directly into it. Your plugin infrastructure, version, and settings remain intact — only the skill files are upgraded.
curl -fsSL https://raw.githubusercontent.com/GadaaLabs/claude-code-on-steroids/main/install.sh | bash
That's it. The next time you open Claude Code, the new skills are live.
What changes: 24 skill files are written into your existing superpowers skills directory. A SessionStart hook is added to ~/.claude/settings.json so your pinned overrides survive any future plugin update.
What stays the same: Your plugin version, all other plugin settings, Claude Code configuration, any custom CLAUDE.md files.
tokenburn CLI (npm install && npm run build && npm link) from the bundled TypeScript source/tokenburn slash commandSessionStart hook that runs apply.sh on every Claude Code session openapply.sh copies 9 pinned skill overrides into the live plugin directory so they survive plugin updatesThe biggest pain point in Claude Code is not missing workflows. It is token burn.
npx claudepluginhub gadaalabs/claude-code-on-steroidsCore skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Harness-native ECC plugin for engineering teams - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.