Garry Tan, CEO of Y Combinator, has ignited a fierce debate in the developer community with “gstack,” a Claude Code configuration that has simultaneously garnered thousands of GitHub stars and intense public backlash since its release in March 2026.
My gstack is now public. It is a set of skills for Claude Code that makes it an agentic engine for building software. https://t.co/P7aOFu5wFM
— Garry Tan (@garrytan) March 12, 2026
The Architecture of the gstack Workflow
Since the project’s launch, Tan has rapidly expanded its capabilities. The gstack GitHub repository currently lists 13 distinct skills, with Tan regularly tweeting about new additions.
The setup functions as a multi-agent system. Tan utilizes one skill to have Claude act as a CEO to evaluate the viability of startup ideas, another for engineering implementation, and a third for security and bug review. These specialized roles extend into design, documentation, and beyond.
The project went viral almost immediately, trending on Product Hunt and accumulating nearly 20,000 stars and over 2,200 forks on GitHub, signaling a massive interest in standardized AI-driven development workflows.
The “God Mode” Controversy
The polarization surrounding the project intensified after Tan shared a text from a CTO friend, who claimed gstack was “god mode” and capable of identifying subtle cross-site scripting (XSS) vulnerabilities that had previously gone unnoticed by his own team.
My CTO friend texted me: “Your gstack is crazy. This is like god mode. Your eng review discovered a subtle cross site scripting attack that I don’t even think my team is aware of. I will make a bet that over 90% of new repos from today forward will use gstack.” https://t.co/P7aOFu5wFM
— Garry Tan (@garrytan) March 12, 2026
The claim drew immediate fire from the tech community. One founder posted to X, arguing that Tan should be embarrassed by the claim and that any CTO relying on such tools for security should be fired. Meanwhile, vlogger Mo Bitar released a critique titled “AI is making CEOs delusional,” labeling the project as little more than “a bunch of prompts” in a text file that experienced developers have already implemented independently.
Sherveen Mashayekhi, founder of Free Agency, echoed these sentiments on Product Hunt, suggesting that the project’s popularity is tied more to Tan’s status as the head of Y Combinator than to the technical novelty of the code itself.
Expert Verdict: Innovation or Overhype?
To cut through the noise, we consulted leading AI models on the efficacy of gstack. The consensus among them is surprisingly favorable regarding the methodology.
ChatGPT assessed that while the prompts aren’t “magical,” the real insight lies in the organizational structure: “AI coding works best when you simulate an engineering org structure, not when you just ask: ‘build this feature.'”
Gemini described the setup as a “Pro” configuration, noting that its primary value is in accuracy rather than just simplifying the coding process. Claude itself weighed in, calling gstack “a mature, opinionated system” and one of the better examples of skill design currently available for its platform.
Reflecting on the intensity of the project’s development, Tan noted in a recent X post: “I love coding but I love coding with AI even more. I speak it listens and we create. I see the structure and it is built. There is no more powerful an experience to me than that.”
