Application Security Weekly

Everyone is interested in generative AIs and LLMs, and everyone is looking for use cases and apps to apply them to. Just as the early days of the web inspired the original OWASP Top 10 over 20 years ago, the experimentation and adoption of LLMs has inspired a Top 10 list of their own. Sandy Dunn talks about why the list looks so familiar in many ways -- after all, LLMs are still software. But the list captures some new concepts that anyone looking to use LLMs or generative AIs should be aware of.

• https://llmtop10.com/
• https://github.com/OWASP/www-project-top-10-for-large-language-model-applications/wiki/Educational-Resources
• https://owasp.org/www-project-ai-security-and-privacy-guide/
• https://gandalf.lakera.ai/
• https://quarkiq.com/blog

How companies are benefiting from the enterprise browser. It's not just security when talking about the enterprise browser. It's the marriage between security AND productivity. In this interview, Mike will provide real live case studies on how different enterprises are benefitting.

Segment Resources:

• https://www.island.io/resources
• https://www.island.io/press

This segment is sponsored by Island. Visit https://www.securityweekly.com/islandrsac to learn more about them!

The cybersecurity landscape continues to transform, with a growing focus on mitigating supply chain vulnerabilities, enforcing data governance, and incorporating AI into security measures. This transformation promises to steer DevSecOps teams toward software development processes with efficiency and security at the forefront. Josh Lemos, Chief Information Security Officer at GitLab will discuss the role of AI in securing software and data supply chains and helping developers work more efficiently while creating more secure code.

This segment is sponsored by GitLab. Visit https://securityweekly.com/gitlabrsac to learn more about them!

Visit https://www.securityweekly.com/asw for all the latest episodes!

A lot of AI security has nothing to do with AI -- things like data privacy, access controls, and identity are concerns for any new software and in many cases AI concerns look more like old-school API concerns. But...there are still important aspects to AI safety and security, from prompt injection to jailbreaking to authenticity. Caleb Sima explains why it's important to understand the different types of AI and the practical tasks necessary to secure how it's used.

Segment resources:

• https://calebsima.com/2023/08/16/demystifing-llms-and-threats/
• https://www.youtube.com/watch?v=qgDtOu17E&t=1s

We already have bug bounties for web apps so it was only a matter of time before we would have bounties for AI-related bugs. Keith Hoodlet shares his experience winning first place in the DOD's inaugural AI bias bounty program. He explains how his education in psychology helped fill in the lack of resources in testing an AI's bias. Then we discuss how organizations should approach the very different concepts of AI security and AI safety.

Segment Resources:

• https://securing.dev/posts/hacking-ai-bias/
• https://www.defense.gov/News/Releases/Release/Article/3659519/cdao-launches-first-dod-ai-bias-bounty-focused-on-unknown-risks-in-llms/

Visit https://www.securityweekly.com/asw for all the latest episodes!

Companies deploy tools (usually lots of tools) to address different threats to supply chain security. Melinda Marks shares some of the chaos those companies still face when trying to prioritize investments, measure risk, and scale their solutions to keep pace with their development. Not only are companies still figuring out supply chain, but now they're bracing for the coming of genAI and how that will just further highlight the current struggles they're having with data security and data privacy.

Misusing random numbers, protecting platforms for code repos and package repos, vulns that teach us about designs and defaults, and more!

Visit https://www.securityweekly.com/asw for all the latest episodes!

How can open source projects find a funding model that works for them? What are the implications with different sources of funding? Simon Bennetts talks about his stewardship of Zed Attack Proxy and its journey from OWASP to OpenSSF to an Open Source Fellowship with Crash Override. Mark Curphy adds how his experience with OWASP and the appsec community motivated him to create Crash Override and help projects like ZAP gain the support they deserve.

Segment resources:

• https://crashoverride.com/blog/welcome-zap-to-the-open-source-fellowship
• https://www.zaproxy.org
• https://crashoverride.com/blog/are-there-too-many-bubbles-of-similar-security-efforts

CISA chimes in on the XZ Utils backdoor, PuTTY's private keys and maintaining a secure design, LeakyCLI and maintaining secure secrets in CSPs, LLMs and exploit generation, and more!

Visit https://www.securityweekly.com/asw for all the latest episodes!

There are as many paths into infosec as there are disciplines within infosec to specialize in. Karan Dwivedi talks about the recent book he and co-author Raaghav Srinivasan wrote about security engineering. There's an appealing future to security taking on engineering roles and creating solutions to problems that orgs face. We talk about the breadth and depth of security engineering and ways to build the skills that will help you in your appsec career.

Segment resources:

• https://kickstartseceng.com

A Rust advisory highlights the perils of parsing and problems of inconsistent approaches, D-Link (sort of) deals with end of life hardware, CSRB recommends practices and processes for Microsoft, Chrome’s V8 Sandbox increases defense, and more!

Visit https://www.securityweekly.com/asw for all the latest episodes!

We look into the supply chain saga of the XZ Utils backdoor. It's a wild story of a carefully planned long con to add malicious code to a commonly used package that many SSH connections rely on. It hits themes from social engineering and abuse of trust to obscuring the changes and suppressing warnings. It also has a few lessons about software development, the social and economic dynamics of open source, and strategies for patching software.

It's an exciting topic partially because so much other appsec is boring. And that boring stuff is important to get right first. We also talk about what parts of this that orgs should be worried about and what types of threats they should be prioritizing instead.

Segment Resources:

• https://tukaani.org/xz-backdoor/
• https://news.risky.biz/risky-biz-news-supply-chain-attack-in-linuxland/
• https://www.zdnet.com/article/this-backdoor-almost-infected-linux-everywhere-the-xz-utils-close-call/#ftag=RSSbaffb68
• https://therecord.media/malicious-backdoor-code-linux-red-hat-cisa
• https://www.cisa.gov/news-events/alerts/2024/03/29/reported-supply-chain-compromise-affecting-xz-utils-data-compression-library-cve-2024-3094
• https://duo.com/decipher/carefully-crafted-campaign-led-to-xz-utils-backdoor
• https://boehs.org/node/everything-i-know-about-the-xz-backdoor

OWASP leaks resumes, defining different types of prompt injection, a secure design example in device-bound sessions, turning an ASVS requirement into practice, Ivanti has its 2000s-era Microsoft moment, HTTP/2 CONTINUATION flood, and more!

Visit https://www.securityweekly.com/asw for all the latest episodes!

Sometimes infosec problems can be summarized succinctly, like "patching is hard". Sometimes a succinct summary sounds convincing, but is based on old data, irrelevant data, or made up data. Adrian Sanabria walks through some of the archeological work he's done to dig up the source of some myths. We talk about some of our favorite (as in most disliked) myths to point out how oversimplified slogans and oversimplified threat models lead to bad advice -- and why bad advice can make users less secure.

Segment resources:

• https://www.oreilly.com/library/view/cybersecurity-myths-and/9780137929214/

The OWASP Top 10 gets its first update after a year, Metasploit gets its first rewrite (but it's still in Perl), PHP adds support for prepared statements, RSA Conference puts passwords on notice while patching remains hard, and more!

Visit https://www.securityweekly.com/asw for all the latest episodes!

One of the biggest failures in appsec is an attitude that blames users for security problems. A lot of processes and workflows break down because of an insecure design or insecure defaults. Benedek Gagyi chats with us about the impact of the user experience (UX) on security and why it's not only important to understand how to make a user's life easier, but in defining who that user is in the first place.

Segment resources:

• https://www.usenix.org/conference/8th-usenix-security-symposium/why-johnny-cant-encrypt-usability-evaluation-pgp-50

The GoFetch side channel in Apple CPUs, OpenSSF's plan for secure software developer education, fuzzing vs. formal verification as a security strategy, hard problems in InfoSec (and AppSec), and more!

Visit https://www.securityweekly.com/asw for all the latest episodes!

Lots of companies need cybersecurity programs, as do non-profits. Tyler Von Moll talks about how to get small organizations started on security and how to prioritize initial investments. While an appsec program likely isn't going to be one of the first steps, it's going to be an early one. What decisions can you make at the start that will benefit the program in the years that follow? What does an appsec program look like at a small scale?

Segment Resources:

• "Cybersecurity for Nonprofits", https://docs.google.com/presentation/d/18HuKtwgwGMtEJ87CgkMqHp1JDVRUXPP--zptjMpF0/edit?usp=sharing
• https://www.verizon.com/business/resources/reports/dbir/2023/master-guide/

Insecure defaults and insecure design in smart locks, FCC adopts Cyber Trust Mark labels for IoT devices, the ZAP project gets a new home, and more!

Visit https://www.securityweekly.com/asw for all the latest episodes!

A majority of internet traffic now originates from APIs, and cybercriminals are taking advantage. Increasingly, APIs are used as a common attack vector because they’re a direct pathway to access sensitive data. In this discussion, Lebin Cheng shares what API attack trends Imperva, a Thales Company has observed over the past year, and what steps organizations can take to protect their APIs.

This segment is sponsored by Imperva. Visit https://www.securityweekly.com/imperva to learn more about them!

The trivial tweaks to bypass authentication in TeamCity, ArtPrompt attacks use ASCII art against LLMs, annoying developers with low quality vuln reports, removing dependencies as part of secure by design, removing overhead with secure by design, and more!

Visit https://www.securityweekly.com/asw for all the latest episodes!