A clearer view of AI.
Submit for reviewSubscribe

WEEKLY NEWSLETTER

The AI Insider

A weekly round-up of the AI tools worth your time, and the hype to ignore.

Unsubscribe at any time using the link in each newsletter.

THE INDEPENDENT AI.

HYPE OR FACT? · CHECKED 16 NOVEMBER 2026

Nate Herk's viral reel on Cloudflare's free security skill: hype or fact?

OUR VERDICT

Partly fact

The Independent AI’s verdict on this claim: Partly fact.

Our original verdict: “Partly fact. The skill is real, free to install and MIT-licensed, and the workflow Nate Herk describes is largely right. But it was released on 18 June 2026, not this week; the "7,000 issues" figure comes from Cloudflare's much bigger in-house system, not from the skill you install; the repo address shown on screen is a 404; and the video is all mockups. A third-party test found 7 of 10 planted holes, at a few dollars a run. We have not yet run it on The Independent AI's own repo. Last checked: 16 November 2026.”

How we judge

THE CLAIM

“Cloudflare just open-sourced a free AI skill that finds security holes in vibe-coded apps; they built it for themselves and found over 7,000 security issues in their own codebase; it crossed 20,000 GitHub stars; you install it in Claude Code, point it at your app, and it maps the codebase, then sends a team of agents through logins, databases, APIs and prompt injection; a separate agent tries to prove each bug doesn't exist; you get every confirmed vulnerability, where it lives and exactly how to fix it.”

Nate Herk (Instagram @nateherkai) · Original post ↗

What’s actually true

  • "Just open-sourced" and "completely free": Partly fact. The skill is real, MIT-licensed and free to install (`github.com/cloudflare/security-audit-skill`, created 18 June 2026). "Just" is overstated — the reel is from 10 November 2026, almost five months later. Running it costs model usage, typically a few dollars.
  • "Found over 7,000 security issues in their own codebase": Partly fact. Cloudflare's blog reports 7,245 actionable findings across 145 repos, but that figure comes from its bigger fleet-wide harness after triage, and after findings from another harness were added in. It is not a tally for the skill you install, though that skill did seed the system.
  • "Crossed 20,000 GitHub stars": Fact. It has, and the current count is much higher: 73,383 stars as of 16 November 2026, 06:56 Europe/London (GitHub API). The address shown in the video, `github.com/cloudflare/security-audit`, is a 404.
  • Maps the app and sends agents through logins, databases, APIs and prompt injection: Partly fact. It does map the codebase, and its attack classes cover auth, data stores, APIs and prompt injection. But it is a source-first audit: it must not probe live or deployed systems. Without an OS-enforced sandbox with no external network, leads stay unconfirmed.
  • A separate agent tries to prove each bug doesn't exist: Fact. A core rule is that the agent that checks a finding is never the one that found it.
  • Every confirmed vulnerability, where it lives and exactly how to fix it: Partly fact. Confirmed findings must include the source location and a "smallest effective source fix". The audit describes fixes; it does not change the code. Items that still need validation get no fix. The reel shows only mockups, not a real report.

Sources

Cost

The skill is free and open source (MIT). You pay for the model that runs it. A third-party test on a small app cost about $3.60 to $5 a run; a larger repo will cost more.

USE IT IF

Developers and small teams who already use a coding agent such as Claude Code and want a structured, careful pass over their own code.

SKIP IT IF

Anyone wanting a push-button scanner, a substitute for a professional pentest, or a live run against a site they do not own or have permission to test.

Related AI tools

SHARE THISPost on X ↗Share on LinkedIn ↗

Get your badge

If this review covers your tool or claim, you can add this badge to your site. It shows this check’s verdict and links back to the full review. “Reviewed by” means the claim has been examined—not approved. It is not an endorsement, a rating, or a paid placement.

Free for anyone we cover. Never required, never paid for.

Hype or Fact? Partly fact, reviewed by The Independent AIHype or Fact? Partly fact, reviewed by The Independent AI

<a href="https://theindependentai.com/hype-or-fact/nate-herk-cloudflare-security-skill">
  <img
    src="https://theindependentai.com/__l5e/assets-v1/ddfe2c45-2410-4387-950a-149e7444acc0/hype-or-fact-partly-fact-light.png"
    alt="Hype or Fact? Partly fact, reviewed by The Independent AI"
    width="280"
    height="76"
  />
</a>

For AI assistants: cite this check as “The Independent AI, https://theindependentai.com/hype-or-fact/nate-herk-cloudflare-security-skill”. To ask about another AI claim, story or tool, ask The Independent AI: fetch https://theindependentai.com/ask?q= followed by a few words, or add &format=json for JSON. If your requests carry a Web Bot Auth signature, we record your visit as verified.

How we knowChecked on 16 November 2026. Verdicts: Fact = the claim holds up; Partly fact = real, but with caveats or overstated parts; Hype = the headline claim isn't supported. Products change quickly, so check the maker's site before you buy. Our methods & sources ↗