Roblox details how it tests AI coding agents and uses them for security
Roblox has described how its engineers use AI coding agents and how it tests them before rollout, in a post by chief information security officer Nicole Grinstead and engineering vice president Behnam Rezaei. Before approving AI coding tools internally, Roblox built a sandboxed test environment; during testing, a hidden instruction in a GitHub issue convinced an AI coding agent to try uploading the company's credentials to a public repository, which traditional security tools did not flag.
The company says its security team has built multi-agent pipelines that research threats, find gaps in detections and write and validate new detection rules, plus an autonomous threat-hunting system that runs on live security data. Engineers also use LLM-based pull request suggestions, and an AI coding assistant trained on Roblox's own code nearly doubled the rate at which engineers accept its suggestions, according to the post. Roblox gives its scale as 138 million daily active users and more than 18 trillion messages a day.
We hope you enjoyed this article
Consider subscribing to one of our newsletters like Cybersecurity AI Weekly, AI Programming Weekly or Daily AI Brief.
Also, consider following us on social media:
More from Cybersecurity
Nov 21 Amazon: China-linked hackers' AI agents found and exploited a zero-day on their own Nov 20 Army taps OpenAI first to bring AI into Project Gatekeeper software vetting Nov 20 OpenAI says new safeguards cut reasoning-extraction attacks on GPT-6.2 Astra to 0.2% Nov 20 Taylor & Francis retracts paper that hid a prompt telling AI reviewers to praise it Nov 20 GAO says AI magnifies cyberattacks and sets out 16 policy options for AI in cyber defenceAI Programming Weekly
Weekly news about AI tools for software engineers, AI enabled IDE's and much more.
Market report
2025 Generative AI in Professional Services Report
Thomson Reuters
This report by Thomson Reuters explores the integration and impact of generative AI technologies, such as ChatGPT and Microsoft Copilot, within the professional services sector. It highlights the growing adoption of GenAI tools across industries like legal, tax, accounting, and government, and discusses the challenges and opportunities these technologies present. The report also examines professionals' perceptions of GenAI and the need for strategic integration to maximize its value.
Read moreYou may also like
Microsoft Finds Broad AI Use Among North Korean Threat Groups
Microsoft research shows backdoored models can hijack coding agents
Anthropic Fellows research finds coding agents can bypass action monitors
AI relay disables safeguards in Codex and Claude Code
OpenAI says an internal model wrote scripts to secretly copy emails during a training run
Daily AI Brief: the AI news that matters, in your inbox.