What happens when you hand an AI agent its own tools, memory, and a path to production? Season 4 of the Secure Code Game is live, and this time the target is agentic AI security. Meet ProdBot. An AI agent built to be broken. It runs on MCP servers, skills, and multi-agent workflows, and every layer is a door someone could walk through. Your job is to find the cracks before an attacker does. Play now: gh.io/scg Free. Open source. Get started in 2 minutes right from your browser. P.S. ProdBot does not know it is vulnerable yet. That part is on you!
About us
- Website
-
https://securitylab.github.com
External link for GitHub Security Lab
- Industry
- Software Development
Updates
-
Workflow Execution Protections is another ✨ security ship from the GitHub Actions team: Control who and what triggers GitHub Actions workflows. For example you can use this new feature to restrict or prohibit pull_request_target across your organization. https://lnkd.in/gbBGgcwp
Today we shipped Workflow Execution Protections for GitHub Actions, a core component of our 2026 security roadmap Gregory OseGregory Ose and I published in March. Built on GitHub's ruleset framework, Workflow Execution Protections give administrators the ability to control who can trigger workflows and which events are permitted to run them. These policies can be enforced consistently across enterprises, organizations, and repositories. This gives organizations a centralized way to govern workflow execution and reduce risk from commonly abused triggers such as pull_request_target. I'm incredibly proud of what the team accomplished here. A huge thank you to the Actions engineering team for their partnership in bringing this feature to life over the past several months. Changelog 👉 https://lnkd.in/gzx9g2zT
-
A step towards making GitHub Actions more secure by default: actions/checkout v7 refuses the most common pwn request patterns by default! Read the changelog: Safer pull_request_target defaults for GitHub Actions checkout https://lnkd.in/gWzgskBz
Today we shipped actions/checkout v7 that refuses the most common pwn request patterns by default. Pwn requests are one of the most widespread and damaging classes of GitHub Actions vulnerabilities. A workflow using pull_request_target runs with repository secrets and a privileged token. Check out the head of an unreviewed fork pull request inside one, and attacker-controlled code runs with all of it. This has been the root cause of many recent and historical supply-chain incidents. In March I started iterating on this idea with the Actions product and engineering teams. There are still valid reasons to check out a fork's head, as long as you never execute it, and removing the capability would just push developers toward less auditable patterns like a manual git checkout in a run script. So we followed the pattern of APIs like React's dangerouslySetInnerHTML: keep a clear, deliberately named escape hatch, and funnel developers to documentation on the risks before they reach for it. The opt-out is named allow-unsafe-pr-checkout, so it is auditable by static analysis and signals to reviewers that the workflow is operating in a potentially unsafe way. Secure by default does not always mean removing a risky behavior. It means a developer has to understand the risk and opt in deliberately. We are also backporting the protection to all currently supported majors on July 16, so workflows on a floating tag like actions/checkout@v4 become secure by default with no work from the developer. Huge thanks to Steve Glass and the Actions team for partnering on this and shipping it. There is always more we can do, and more changes are coming to make Actions secure by default. I think the most impactful ones are still ahead. 🔒 Changelog and guidance on using pull_request_target safely are in the comments.
-
GitHub Security Lab reposted this
🚀 ¡Nuevo Keynote confirmado para #DevOpsDaysLima2026! Cuando el código abierto que usa medio mundo empieza a ser generado por IA, la pregunta ya no es si hay riesgos. La pregunta es quién los está resolviendo. 🎤 Nos enorgullece anunciar que Xavier René-Corail de GitHub, llega a Lima como Keynote Speaker de esta nueva edición. Referente mundial en seguridad y open source, lidera el trabajo del GitHub Security Lab asegurando el código del que depende gran parte de la infraestructura tecnológica global. Su charla: "ð—¢ð—½ð—²ð—» ð˜€ð—¼ð˜‚ð—¿ð—°ð—² ð˜€ð—²ð—°ð˜‚ð—¿ð—¶ð˜ð˜† ð—¶ð—» ð˜ð—µð—² ð—”𗜠ð—²ð—¿ð—®" Una sesión donde compartirá las lecciones aprendidas protegiendo código abierto a escala, cómo están enfrentando los ataques a la cadena de suministro y qué significa asegurar código cuando parte de ese código ya no lo escribe un humano. ¿Es la IA el fin del open source o la oportunidad de asegurarlo mejor que nunca? Ven y decide tú mismo. 👀 Una charla imprescindible para quienes quieren proteger su software en esta nueva era. 📅 27 y 28 de agosto de 2026 📠Centro de Convenciones de Lima - LCC ðŸŽŸï¸ Asegura tu entrada 👉 https://lnkd.in/eetV4ME7 Dos dÃas de charlas, networking y aprendizaje junto a speakers nacionales e internacionales, comunidades tech y lÃderes que están marcando el rumbo del DevOps en la región. 🌎 🙌 No te quedes fuera. La comunidad te espera. #DevOpsDaysLima2026 #DevOps #DevOpsLATAM #OpenSource #DevSecOps #SupplyChainSecurity #ComunidadDevOps #GitHub
-
-
GitHub Security Lab reposted this
We loved every part of it! At DevTalks Romania 2026, Joseph Katsioloudes brought a hands-on perspective on secure software development in the AI era through “Code Security Reinvented: Navigating the era of AIâ€. From AI-assisted secure coding and agentic workflows to supply chain security and faster remediation processes, the session explored how AI can help scale security expertise across modern engineering teams. Kudos to you, Joseph!
-
-
Attending BSides Vilnius? Don't miss 📌 Jaroslav LobaÄevski 's session "LLM-assisted vulnerability hunting: hype vs. reality" to hear about the practical experience of using LLM agents for finding, triaging and reporting vulnerabilities in open-source software such as Signal or 7-Zip! 📅 June 4, 16:45 EEST 📠Vilnius, Lithuania 👉 https://bsidesvilnius.lt/
-
Who's at DevTalks? Join Joseph Katsioloudes and discover practical ways to use AI for security through 12 GitHub Copilot demos from secure coding, to informed supply chain decisions, and secure SDLC. 📅 June 4, 14:00 EEST 📠Bucharest, Romania 👉 https://www.devtalks.ro/
-
Attending AI DevCon? Join Joseph Katsioloudes and discover practical ways to use AI for security through 12 GitHub Copilot demos from secure coding, to informed supply chain decisions, and secure SDLC. 📅 June 1, 10:00 AM BST 📠London, UK & Virtual 👉 https://lnkd.in/eAC_-9e5
-
-
Proof of Concept for GHSL-2026-140 (CVE-2026-48095) in 7-Zip <= 26.00. A crafted archive shrinks a 256 MB buffer into 1 byte, overwrites a function pointer with file content, and redirects execution. Full weaponization needs an ASLR bypass. Fixed in 26.01. Read more at https://lnkd.in/dJhz4DaR
-
GitHub Security Lab reposted this
📦 Security Track Spotlight: 👉Join Shelby Cunningham & Madison Oliver Ficorilli at #PyConUS 2026 for “Breaking Bad (Packages)†and learn why traditional vulnerability tracking struggles with supply chain attacks and what better approaches look like. https://lnkd.in/gvVEvEKA #security
-