garrytan/gstack

Use Garry Tan's exact Claude Code setup: 23 opinionated tools that serve as CEO, Designer, Eng Manager, Release Manager, Doc Engineer, and QA

What it solves

gstack transforms a single AI coding assistant (like Claude Code) into a virtual engineering team. It solves the problem of "AI slop" and the lack of structured product thinking by replacing blank prompts with a rigorous, multi-role software development lifecycle (SDLC) process. Instead of just generating code, it forces the user through a disciplined pipeline of product interrogation, architectural planning, and quality assurance.

How it works

It operates as a set of "skills" (slash commands) that can be installed into various AI agents (Claude Code, OpenAI Codex CLI, Cursor, etc.). These skills are designed to be run in a specific sequence—Think, Plan, Build, Review, Test, and Ship—where the output of one stage (e.g., a design doc from /office-hours) serves as the input for the next (e.g., /plan-ceo-review).

It provides specialized personas to handle different aspects of development:

  • Product/CEO: Reframes the problem and challenges scope.
  • Engineering Manager: Locks architecture and data flow.
  • Designer: Evaluates design dimensions and catches visual errors.
  • QA/Security: Runs browser-based tests and security audits (OWASP/STRIDE).
  • Release Engineer: Manages the final PR and deployment.

Who it’s for

  • Technical Founders and CEOs who want to ship production-ready products rapidly without a full human team.
  • First-time Claude Code users who need structured roles rather than a blank prompt.
  • Tech leads and staff engineers seeking rigorous review, QA, and release automation for every pull request.

Highlights

  • Virtual Team Personas: Includes 23 specialists and 8 power tools as slash commands.
  • Cross-Agent Support: Compatible with Claude Code, OpenAI Codex CLI, Cursor, GitHub Copilot CLI, and others.
  • Integrated Browser Skills: Uses the Aside browser or a bundled Chromium instance for real-world QA and scraping.
  • End-to-End Workflow: A structured pipeline from /office-hours (product discovery) to /ship (deployment).
  • Security Auditing: Includes a /cso skill for running security audits.

Related

  • Project
    backnotprop/pstack

    A collection of agent skills and playbooks that enforce rigorous engineering principles on AI coding agents to reduce code bloat and increase quality.

  • Project
    echoVic/boss-skill

    An auditable agent-team workflow for coding agents that turns a single agent into a structured engineering team with verifiable quality gates and event-sourced state.

  • Project
    rsmdt/the-startup

    A multi-agent AI framework for Claude Code that implements a spec-driven development workflow to turn ideas into shipped code via structured planning and parallel specialist agents.

  • Project
    deepset-ai/haystack

    An open-source AI orchestration framework for building production-ready RAG systems, autonomous agents, and LLM applications in Python.

  • Dispatch
    Command Center: An Agentic Coding Environment for Production Quality

    Command Center is an agentic coding environment designed to reduce the time spent reviewing AI-generated code by providing structured walkthroughs and refactoring agents to ensure production-grade quality.