Production-grade cross-platform engineering plugin for Claude Code, Cowork, and Codex. 17 discipline skills (frontend, backend, full-stack, mobile, game, devops, API, database, embedded, cloud, AI/ML, blockchain, QA, security, AR/VR, data science, web). 4 cross-cutting agents (discipline-router, build-validator, code-reviewer, ship-auditor). 18 slash commands (one per discipline plus a /uber router). 21 lifecycle hooks (PostToolUse skill validators, PreToolUse destructive-action guards, Stop ship-auditor reminder). MCP wiring for 9 platforms: Supabase, GitHub, Cloudflare, Stripe, PostHog, Context7, Playwright, Chrome DevTools, Firecrawl.
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
AI/ML Development — invoke the ai-ml-development skill with focused intent.
API Development — invoke the api-development skill with focused intent.
Backend Development — invoke the backend-development skill with focused intent.
Blockchain Development — invoke the blockchain-development skill with focused intent.
Cloud Development — invoke the cloud-development skill with focused intent.
Verifies that a claimed change actually works end-to-end before any "done" claim ships to the user. Use proactively after any edit, deploy, or migration. This agent enforces the user's "Definition of Done" rule: build success ≠ done; visible correct output = done.
Reviews a diff or PR for correctness, security, performance, and maintainability before merge. Use after writing or modifying code. Reports only high-confidence, high-impact issues — not nitpicks.
Routes ambiguous or cross-cutting engineering requests to the right discipline skill(s) inside the uber-engineer plugin. Use when the user asks for something that touches multiple layers (e.g. "ship a feature with UI + API + DB + tests") or when the right discipline isn't obvious from the request.
Audits "done" claims against actual visible output. Use proactively when an agent or human says "shipped", "deployed", "verified", "done", or claims a user can now see something. Catches false completions where build/deploy succeeded but the visible surface is broken.
Experiment workflows, datasets, evals, model packaging, serving, and rollback for AI/ML systems. Use when the user mentions: ML, machine learning, AI, model training, fine-tuning, PyTorch, TensorFlow, JAX, scikit-learn, Hugging Face, LangChain, LlamaIndex, evals, RAG, vector database, embeddings, MLflow, Weights & Biases, prompt engineering, Anthropic API, OpenAI API. Pair with the discipline-router agent for cross-cutting work. Do NOT trigger for: data analysis without model training (use data-science-development); ML infra rollout work without model changes (use devops-and-infrastructure).
Public-facing API contract design: paths, versioning, backward compatibility, error schemas, webhook signing, SDK generation, developer experience. Use when the user mentions: API design, REST design, GraphQL schema, gRPC proto, OpenAPI, Swagger, API versioning, deprecation policy, error schema, RFC 7807, problem details, webhook contract, cursor pagination, API key issuance, OAuth scopes, SDK generation, API gateway, public docs. Pair with the discipline-router agent for cross-cutting work. Do NOT trigger for: implementing the handlers behind the contract (use backend-development — this skill designs the wire; backend-development builds the service); internal-only service-to-service calls without external contract; frontend data fetching strategy without server contract change.
Spatial UX, perf budgets, motion comfort, anchors, hand/eye input, and platform-specific deploy. Use when the user mentions: AR, VR, XR, spatial, Vision Pro, visionOS, Quest, Meta Horizon, ARKit, ARCore, WebXR, OpenXR, Unity XR, Unreal XR, RealityKit, RealityComposer, hand tracking, eye tracking, passthrough, motion sickness. Pair with the discipline-router agent for cross-cutting work. Do NOT trigger for: flat 2D UI work (use frontend-development); mobile non-XR apps (use mobile-development).
Service implementation behind the wire contract: handlers, auth enforcement, persistence integration, idempotency, caching, queues, observability. Use when the user mentions: backend, server, service, handler, Node.js, Python, Go, Java, Spring Boot, Express, Fastify, FastAPI, Django, Rails, authentication enforcement, authorization, JWT verification, OAuth callback, session, rate limiting, idempotency, caching, queue, background job, worker. Pair with the discipline-router agent for cross-cutting work. Do NOT trigger for: public API contract design / versioning / SDK generation (use api-development — this skill implements behind the contract; api-development designs the contract); pure frontend UI work (use frontend-development); database schema migration work (use database-development); container orchestration / deploy pipeline (use devops-and-infrastructure).
Smart contracts, wallet UX, indexers, security review, and on-chain integration. Use when the user mentions: blockchain, Ethereum, Solidity, Foundry, Hardhat, Rust smart contracts, Solana, Anchor, wallet, ERC-20, ERC-721, NFT, DeFi, EVM, subgraph, indexer, RPC, MEV, wagmi, viem, ethers.js. Pair with the discipline-router agent for cross-cutting work. Do NOT trigger for: general web3 frontend without contract code (use frontend-development); trading bots without on-chain integration.
Supabase project management — schemas, migrations, RLS, edge functions, branches.
GitHub repos, PRs, issues, actions, gists.
Cloudflare Workers, D1, KV, R2, Pages, Hyperdrive, Durable Objects.
Stripe payments, customers, subscriptions, invoices, webhooks.
PostHog analytics, feature flags, experiments, session replays, error tracking.
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Requires secrets
Needs API keys or credentials to function
Uses power tools
Uses Bash, Write, or Edit tools
Production-grade engineering plugin for Claude Code, Cowork, and Codex. 17 discipline skills, 4 specialized agents, MCP wiring for the modern web stack, and a deterministic "definition of done" workflow.
Most plugins ship one capability. Real engineering work crosses disciplines: a feature touches UI, API, database, tests, and deploy. This plugin gives an agent the right vocabulary, official-doc references, anti-pattern catalog, and verification rubric for every layer in one installable unit.
The doctrine: build success ≠ done. 200 OK ≠ done. A real user seeing correct output = done.
# Clone the marketplace
git clone https://github.com/Blaze-sports-Intel/uber-engineer.git ~/.claude/marketplaces/uber-engineer
# Inside Claude Code
/plugin marketplace add ~/.claude/marketplaces/uber-engineer
/plugin install uber-engineer@uber-engineer-marketplace
Or one-liner:
curl -fsSL https://raw.githubusercontent.com/Blaze-sports-Intel/uber-engineer/main/install.sh | bash
The same plugins/uber-engineer/ directory is Cowork-compatible. Drop it into your Cowork plugins
directory or install via the Cowork plugin loader UI.
codex-cli 0.120+ has no plugin install subcommand. Codex integration uses AGENTS.md plus
symlinked skills and the codex mcp add flow:
git clone https://github.com/Blaze-sports-Intel/uber-engineer.git
cd uber-engineer/codex-mirror/uber-engineer
mkdir -p ~/.codex/skills ~/.codex/plugins
ln -sfn "$(pwd)" ~/.codex/plugins/uber-engineer
ln -sfn "$(pwd)/skills/"* ~/.codex/skills/
cp AGENTS.md ~/.codex/AGENTS.md # or merge with an existing one
Wire the MCP servers Codex should call (env vars must be set first):
codex mcp add github -- npx -y @modelcontextprotocol/server-github
codex mcp add cloudflare -- npx -y @cloudflare/mcp-server-cloudflare
# ...repeat for each MCP server you want Codex to reach
| # | Skill | Command | What it covers |
|---|---|---|---|
| 01 | frontend-development | /uber:frontend | UI implementation, design systems, accessibility, responsive layout, and frontend performance. |
| 02 | backend-development | /uber:backend | Service architecture, API contracts, auth, data persistence, and operational hygiene. |
| 03 | full-stack-development | /uber:fullstack | End-to-end feature delivery across frontend, backend, API, database, and deploy. |
| 04 | mobile-development | /uber:mobile | Native iOS, native Android, and cross-platform mobile with build, test, store, and offline constraints. |
| 05 | game-development | /uber:game | Engine architecture, gameplay systems, performance budgets, and asset pipelines. |
| 06 | devops-and-infrastructure | /uber:devops | CI/CD, infrastructure as code, observability, rollback, and incident hygiene. |
| 07 | api-development | /uber:api | API contracts, versioning, backward compatibility, error schemas, and developer experience. |
| 08 | database-development | /uber:db | Schema design, migration safety, indexing, query review, and read-only-by-default access. |
| 09 | embedded-systems-development | /uber:embedded | C/C++/Rust on microcontrollers, RTOS, memory limits, hardware abstraction, and cross-compiling. |
| 10 | cloud-development | /uber:cloud | Serverless, multi-environment config, observability, and release promotion across major clouds. |
| 11 | ai-ml-development | /uber:ai | Experiment workflows, datasets, evals, model packaging, serving, and rollback for AI/ML systems. |
| 12 | blockchain-development | /uber:blockchain | Smart contracts, wallet UX, indexers, security review, and on-chain integration. |
| 13 | test-and-quality-assurance | /uber:qa | Test strategy, contract tests, E2E, visual regression, performance, and accessibility verification. |
| 14 | security-development | /uber:security | Threat modeling, secure-by-default code, dependency hygiene, secrets management, and incident response. |
| 15 | ar-vr-development | /uber:xr | Spatial UX, perf budgets, motion comfort, anchors, hand/eye input, and platform-specific deploy. |
| 16 | data-science-development | /uber:data | Reproducible analysis, dataset hygiene, statistical rigor, dashboarding, and shipping insights. |
| 17 | web-development | /uber:web | End-to-end web app delivery: routing, rendering modes, SEO, perf, deploy, and observability. |
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 claimnpx claudepluginhub blaze-sports-intel/uber-engineer --plugin uber-engineerTwo-skill engineering pack: api-craft (wire contract design — REST/GraphQL/RPC choice, versioning, error envelopes, pagination, idempotency, rate limiting, webhooks, SDK ergonomics, OpenAPI as source-of-truth) + mobile-craft (mobile orchestration and decision-making — stack-choice phase, cross-runtime patterns for offline/push/deep-linking, cross-store shipping discipline). Sits above implementation specialists: hands off API handlers to fullstack-engineer / cloudflare, hands off Apple-platform work to ios-craft, RN to react-native-foundations, Expo to expo. Includes 2 cross-cutting agents (api-contract-reviewer, mobile-stack-decider) and 7 slash commands.
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.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Access thousands of AI prompts and skills directly in your AI coding assistant. Search prompts, discover skills, save your own, and improve prompts with AI.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
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.
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.