STARTUP AGENTS Skills Hub & Engine
Lensetek Standard | AI tools for startup teams

Orchestrate a Virtual Startup Team in Seconds

Connect 26 specialized roles with structured boundaries, input-output specifications, and strict compliance rules to automate business strategies, development pipelines, and launches.

Windows Package (.zip) Windows Script (.bat) Linux / macOS (.sh) Open on Hugging Face
Startup Agent Skills Hub - Virtual AI Team for Startup Builders | Product Hunt
Method 1: Install by Prompt

Copy and send this prompt to your AI assistant to install the agent skills suite globally:

AI Assistant Prompt: Recommended
Please install the Startup-Agent-Skills-Hub suite globally across my environment from https://huggingface.co/spaces/lensetek/startup-agent-skills-hub so I can use all skills across all my workspaces and projects.
Global Execution Command: npx --yes skills add -g https://huggingface.co/spaces/lensetek/startup-agent-skills-hub.git -y
Method 2: Installation Scripts

Select a package or script matching your operating system and inspect the script before running:

📦 Windows Package (.zip) 🪟 Windows Script (.bat) 🐧 Linux / macOS (.sh)
// Terminal One-Liner (macOS & Linux):
curl -fsSL https://huggingface.co/spaces/lensetek/startup-agent-skills-hub/resolve/main/installers/install.sh | bash
AI Desktop Clients (Optional)
GUI Workspace

Enhance your agent experience with modern desktop graphical workspaces featuring visual diffs, multi-model switching, and project management:

Freebuff Desktop

Universal AI Agent Workspace for Windows.

Download Freebuff

OpenCode (x64)

For standard Intel / AMD laptops & PCs.

Download x64

OpenCode (ARM64)

For Snapdragon X Elite / Copilot+ PCs.

Download ARM64
CLI Installation (Advanced)

If you prefer using the command line interface, you can instantly install all agent modules locally or globally using NPX.

// Install to current project:
npx skills add https://huggingface.co/spaces/lensetek/startup-agent-skills-hub.git
// Install globally (Multi-Project):
npx skills add -g https://huggingface.co/spaces/lensetek/startup-agent-skills-hub.git
Compatible AI Platforms

Supported AI Coding Agents

Integrate our skill modules directly into various leading AI Coding Agents in the market today.

Google logo
Google

Antigravity / Gemini

Autonomous AI coding assistant from Google DeepMind designed specifically to understand large codebases, pair program interactively, and execute complex tools.

Visit Site
Anthropic logo
Anthropic

Claude Code

Autonomous agentic coding CLI by Anthropic that works directly in the terminal to edit files, run tests, and intelligently resolve complex bugs.

Visit Site
OpenAI logo
OpenAI

OpenAI Codex

AI coding engine from OpenAI that translates natural language instructions into functional software systems and interactive command execution.

Visit Site
GitHub Copilot logo
GitHub

GitHub Copilot

World-leading AI developer tool embedded natively in Visual Studio Code, providing context-aware multi-file completions and chat assistance.

Visit Site
Universal

Cursor / Windsurf

Next-generation AI-first code editors built on VS Code foundation with deep repository indexing and autonomous agent modes.

Visit Site
Structured Execution

Startup Delivery Workflows

Follow battle-tested workflows designed to transform ideas into audited, production-ready code.

PHASE 1

Strategy & Scoping

CEO defines vision, Market Analyst inspects competitors with REA, and PM bounds MVP features with MoSCoW prioritization.

PHASE 2

Design & Architecture

UI/UX creates wireframes & contrast guidelines, Scrum Master structures sprint tickets, and PRD is finalized.

PHASE 3

Engineering & E2E QA

Frontend/Backend code, QA runs Playwright cross-browser tests (WebKit/Firefox), network mocking, and records bug video proof.

PHASE 4

Audit & Launch

Git Guard scans for leaks, REA audits decompile security, CI/CD deploys staging, and Marketing executes 30-day campaign.

Modular Architecture

26 Specialized Agent Skills

Every agent has explicit responsibilities, strict boundaries, and standardized input/output contracts.

Tailored Experience

Personalize Agents to Your Stack

Make agents smart and highly personalized by defining your project DNA. AI agents will automatically adapt their outputs to match your custom libraries, security requirements, and target customer personas without needing constant reminders.

// Copy and send this prompt to personalize your agents automatically
Analyze my project structure and dependencies, then automatically generate my startup_dna.md file inside the .agents directory.
Generated: .agents/startup_dna.md
# Startup DNA & Project Preferences
## 1. Business Profile
- Project Name: CleanEat
- Target Market: Jakarta urban professionals
- Primary Language: English (Marketing, docs, and code)
## 2. Detected Tech Stack
- Frontend: Vanilla HTML/JS
- Backend: Node.js (Express)
- Styling: Vanilla CSS
## 3. Security Rules
- API Keys must be stored in .env
- Mobile-first prioritization

Strict Credential Security Policy

SECURITY STANDARD VERIFIED
Mandatory Policy

Server-Side Storage Boundary

All sensitive configurations, tokens, database passwords, and API keys (e.g. Stripe, Gemini) must remain encapsulated in server environments, accessed strictly through `.env` variable lookups.

Zero Client-Side Keys

Frontend templates and client-side binary distributions are prohibited from direct external credentialed calls. All requests must route through local security proxy handlers.

RLS & Firebase Guards

PostgreSQL Row Level Security (RLS) and custom authentication matching on Firebase rules prevent unauthorized multi-tenant or generic wildcard access to database schemas.

DevOps Pre-Push Scan

The `devops-git-guard` module automatically inspects repository changes, audits stage directories for accidentally saved keys, and validates `.gitignore` paths before sync commits.
FAQ

Startup Agent Skills Hub FAQ

Quick answers for search engines, AI answer engines, and builders evaluating the skills hub.

What is Startup Agent Skills Hub?

Startup Agent Skills Hub is a collection of 26 modular AI agent skill definitions for orchestrating startup strategy, product planning, design, engineering, QA, DevOps, computer-use UI automation, Chrome DevTools & Playwright dual-engine browser automation, Windows-MCP native desktop automation, REA deep reverse engineering, DevStudio MCP video recordings, app tutorial manual generation, Graphify AST Knowledge Graphing, marketing, compliance, Obsidian PKM architecture, and launch workflows.

Which AI coding agents are supported?

The skills are designed for AI coding workflows including Google Antigravity and Gemini, Claude Code, Codex, GitHub Copilot, and custom multi-agent frameworks.

How does Playwright MCP integrate with the skills hub?

Playwright MCP provides official browser automation by Microsoft across Chromium, Firefox, and WebKit (Safari). It connects with developer-qa-reviewer for cross-browser testing, network response and payment mocking (browser_route), accessibility ref-based snapshot navigation, and automated CI/CD test generation (@playwright/test).

How does REA (Reverse Engineer Anything) integrate with the skills hub?

REA (rea.tools) provides evidence-based reverse engineering. It empowers market-analyst to inspect competitor binaries, Electron .asar bundles, and mobile APKs to reveal undocumented API endpoints and architecture. Additionally, developer-qa-reviewer uses REA for pre-release decompilation security audits to guarantee no secret tokens or IP leak in production builds.

How does DevStudio MCP integrate with the skills hub?

DevStudio MCP provides programmatic screen recording with bundled PyAV FFmpeg for H.264/AAC MP4 video capture. It connects to app-tutorial-manual-generator for video walkthrough demonstrations and developer-qa-reviewer for capturing definitive bug reproduction video clips.

How does the project protect credentials?

The project requires secrets to stay server-side, routes credentialed calls through backend proxies, separates public client config from private keys, and includes QA and Git Guard checks for exposed credentials.