AI slop is breaking bug bounty programs

Open original

Google paused its open-source bug bounty after AI flooded it with fake vulnerability reports, forcing security teams to waste time triaging hallucinated exploits. Nulsec proposes an automated sandbox that pre-screens submissions to filter out AI slop before human engineers see them.

  • Google froze its open-source bug bounty program because AI-generated reports created an unsustainable triage load—most submissions describe non-existent vulnerabilities or non-reproducible exploits, wasting hundreds of engineering hours.
  • Nulsec's core mechanism: automatically execute reproduction tests in isolated sandboxes for each incoming report, blocking hallucinated claims before they reach human security engineers.
  • The business model stacks three revenue streams (tiered SaaS, usage-based compute, enterprise integrations) and uses free tier access to top GitHub repos as a go-to-market wedge.

Google shut down its open-source bug bounty program because AI-generated vulnerability reports flooded the system faster than human engineers could evaluate them. These submissions sound technical and convincing but describe non-existent vulnerabilities or impossible exploits. Security teams waste hundreds of hours trying to reproduce hallucinated flaws, which pulls focus away from actual critical bugs. The noise-to-signal ratio has become unsustainable across the industry. Nulsec proposes an automated screening layer that sits between vulnerability reporters and security teams. When someone submits a bug report, Nulsec parses the claim and runs dynamic reproduction tests in isolated sandboxes to verify whether the vulnerability actually exists. If the report relies on hallucinated code or can't be reproduced, it gets filtered out or flagged for proof-of-concept verification before any human engineer touches it. The company plans to monetize through tiered SaaS subscriptions ($299–$2,499/month), usage-based compute credits for sandbox execution, and custom enterprise integrations. The business strategy leans on free tools to build market share—releasing an open-source GitHub Action for maintainers and offering free triage sandboxing to the top 1,000 GitHub repositories. They'd also publish a "Slop Index" dashboard tracking AI-generated exploit hallucinations in real time. The competitive advantage sits in a proprietary dataset of hallucinated exploit patterns and execution logs gathered across thousands of repositories. As LLMs keep generating new variations of phantom bugs, Nulsec's sandbox logs create a growing negative training set that competitors would struggle to replicate.

Ask about this article