Application Security Weekly (Video)

A smaller attack surface should lead to a smaller list of CVEs to track, which in turn should lead to a smaller set of vulns that you should care about. But in practice, keeping something like a container image small has a lot of challenges in terms of what should be considered minimal. Neil Carpenter shares advice and anecdotes on what it takes to refine a container image and to change an org's expectations that every CVE needs to be fixed. Show Notes: https://securityweekly.com/asw-344

Open source software is a massive contribution that provides everything from foundational frameworks to tiny single-purpose libraries. We walk through the dimensions of trust and provenance in the software supply chain with Janet Worthington. And we discuss how even with new code generated by LLMs and new terms like slopsquatting, a lot of the most effective solutions are old techniques. Resources https://www.forrester.com/blogs/make-no-mistake-software-is-a-supply-chain-and-its-under-attack/ https://www.forrester.com/report/the-future-of-software-supply-chain-security/RES184050 Show Notes: https://securityweekly.com/asw-343

Maintaining code is a lot more than keeping dependencies up to date. It involved everything from keeping old code running to changing frameworks to even changing implementation languages. Jonathan Schneider talks about the engineering considerations of refactoring and rewriting code, why code maintenance is important to appsec, and how to build confidence that adding automation to a migration results in code that has the same workflows as before. Resources https://docs.openrewrite.org https://github.com/openrewrite Then, instead of our usual news segment, we do a deep dive on some recent vulns NVIDIA's Triton Inference Server disclosed by Trail of Bits' Will Vandevanter. Will talks about the thought process and tools that go into identify potential vulns, the analysis in determining whether they're exploitable, and the disclosure process with vendors. He makes the important point that even if something doesn't turn out to be a vuln, there's still benefit to the learning process and gaining experience in seeing the different ways that devs design software. Of course, it's also more fun when you find an exploitable vuln -- which Will did here! Resources https://nvidia.custhelp.com/app/answers/detail/a_id/5687 https://github.com/triton-inference-server/server https://blog.trailofbits.com/2025/07/31/hijacking-multi-agent-systems-in-your-pajamas/ https://blog.trailofbits.com/2025/07/28/we-built-the-security-layer-mcp-always-needed/ Show Notes: https://securityweekly.com/asw-342

A successful strategy in appsec is to build platforms with defaults and designs that ease the burden of security choices for developers. But there's an important difference between expecting (or requiring!) developers to use a platform and building a platform that developers embrace. Julia Knecht shares her experience in building platforms with an attention to developer needs, developer experience, and security requirements. She brings attention to the product management skills and feedback loops that make paved roads successful -- as well as the areas where developers may still need or choose their own alternatives. After all, the impact of a paved road isn't in its creation, it's in its adoption. Show Notes: https://securityweekly.com/asw-341

AI is more than LLMs. Machine learning algorithms have been part of infosec solutions for a long time. For appsec practitioners, a key concern is always going to be how to evaluate the security of software or a system. In some cases, it doesn't matter if a human or an LLM generated code -- the code needs to be reviewed for common flaws and design problems. But the creation of MCP servers and LLM-based agents is also adding a concern about what an unattended or autonomous piece of software is doing. Sohrob Kazerounian gives us context on how LLMs are designed, what to expect from them, and where they pose risk and reward to modern software engineering. Resources https://www.vectra.ai/research Show Notes: https://securityweekly.com/asw-340

What are some appsec basics? There's no monolithic appsec role. Broadly speaking, appsec tends to branch into engineering or compliance paths, each with different areas of focus despite having shared vocabularies and the (hopefully!) shared goal of protecting software, data, and users. The better question is, "What do you want to secure?" We discuss the Cybersecurity Skills Framework put together by the OpenSSF and the Linux Foundation and how you might prepare for one of its job families. The important basics aren't about memorizing lists or technical details, but demonstrating experience in working with technologies, understanding how they can fail, and being able to express concerns, recommendations, and curiosity about their security properties. Resources: https://cybersecurityframework.io https://owasp.org/www-project-cheat-sheets/ https://blog.cloudflare.com/rfc-8446-aka-tls-1-3/ https://aflplus.plus/ https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-doing-and-why-does-it-work/ Show Notes: https://securityweekly.com/asw-339

Appsec still deals with ancient vulns like SQL injection and XSS. And now LLMs are generating code along side humans. Sandy Carielli and Janet Worthington join us once again to discuss what all this new code means for appsec practices. On a positive note, the prevalence of those ancient vulns seems to be diminishing, but the rising use of LLMs is expanding a new (but not very different) attack surface. We look at where orgs are investing in appsec, who appsec teams are collaborating with, and whether we need security awareness training for LLMs. Resources: https://www.forrester.com/blogs/application-security-2025-yes-ai-just-made-it-harder-to-do-this-right/ Show Notes: https://securityweekly.com/asw-338

Manual secure code reviews can be tedious and time intensive if you're just going through checklists. There's plenty of room for linters and compilers and all the grep-like tools to find flaws. Louis Nyffenegger describes the steps of a successful code review process. It's a process that starts with understanding code, which can even benefit from an LLM assistant, and then applies that understanding to a search for developer patterns that lead to common mistakes like mishandling data, not enforcing a control flow, or not defending against unexpected application states. He explains how finding those kinds of more impactful bugs are rewarding for the reviewer and valuable to the code owner. It involves reading a lot of code, but Louis offers tips on how to keep notes, keep an app's context in mind, and keep code secure. Segment Resources: https://pentesterlab.com/live-training/ https://pentesterlab.com/appsecschool https://deepwiki.com https://daniel.haxx.se/blog/2025/05/29/decomplexification/ Show Notes: https://securityweekly.com/asw-337

Fuzzing has been one of the most successful ways to improve software quality. And it demonstrates how improving software quality improves security. Artur Cygan shares his experience in building and applying fuzzers to barcode scanners, smart contracts, and just about any code you can imagine. We go through the useful relationship between unit tests and fuzzing coverage, nudging fuzzers into deeper code paths, and how LLMs can help guide a fuzzer into using better inputs for its testing. Resources https://blog.trailofbits.com/2024/10/31/fuzzing-between-the-lines-in-popular-barcode-software/ https://github.com/crytic/echidna https://github.com/crytic/medusa https://lcamtuf.blogspot.com/2014/11/pulling-jpegs-out-of-thin-air.html Show Notes: https://securityweekly.com/asw-336

What makes a threat modeling process effective? Do you need a long list of threat actors? Do you need a long list of terms? What about a short list like STRIDE? Has an effective process ever come out of a list? Farshad Abasi joins our discussion as we explain why the answer to most of those questions is No and describe the kinds of approaches that are more conducive to useful threat models. Resources: https://www.eurekadevsecops.com/agile-devops-and-the-threat-modeling-disconnect-bridging-the-gap-with-developer-insights/ https://www.threatmodelingmanifesto.org https://kellyshortridge.com/blog/posts/security-decision-trees-with-graphviz/ In the news, learning from outage postmortems, an EchoLeak image speaks a 1,000 words from Microsoft 365 Copilot, TokenBreak attack targets tokenizing techniques, Google's layered strategy against prompt injection looks like a lot like defending against XSS, learning about code security from CodeAuditor CTF, and more! Show Notes: https://securityweekly.com/asw-335

CISA has been championing Secure by Design principles. Many of the principles are universal, like adopting MFA and having opinionated defaults that reduce the need for hardening guides. Matthew Rogers talks about how the approach to Secure by Design has to be tailored for Operational Technology (OT) systems. These systems have strict requirements on safety and many of them rely on protocols that are four (or more!) decades old. He explains how the considerations in this space go far beyond just memory safety concerns. Segment Resources: https://www.cisa.gov/sites/default/files/2025-01/joint-guide-secure-by-demand-priority-considerations-for-ot-owners-and-operators-508c_0.pdf https://www.youtube.com/watch?v=vHSXu1P4ZTo Show Notes: https://securityweekly.com/asw-334

The recent popularity of MCPs is surpassed only by the recent examples deficiencies of their secure design. The most obvious challenge is how MCPs, and many more general LLM use cases, have erased two decades of security principles behind separating code and data. We take a look at how developers are using LLMs to generate code and continue our search for where LLMs are providing value to appsec. We also consider what indicators we'd look for as signs of success. For example, are LLMs driving useful commits to overburdened open source developers? Are LLMs climbing the ranks of bug bounty platforms? In the news, more examples of prompt injection techniques against LLM features in GitLab and GitHub, the value (and tradeoffs) in rewriting code, secure design lessons from a history of iOS exploitation, checking for all the ways to root, and NIST's approach to (maybe) measuring likely exploited vulns. Show Notes: https://securityweekly.com/asw-333

ArmorCode unveils Anya—the first agentic AI virtual security champion designed specifically for AppSec and product security teams. Anya brings together conversation and context to help AppSec, developers and security teams cut through the noise, prioritize risks, and make faster, smarter decisions across code, cloud, and infrastructure. Built into the ArmorCode ASPM Platform and backed by 25B findings, 285+ integrations, natural language intelligence, and role-aware insights, Anya turns complexity into clarity, helping teams scale securely and close the security skills gap. Anya is now generally available and included as part of the ArmorCode ASPM Platform. Visit https://securityweekly.com/armorcodersac to request a demo! As 'vibe coding", the practice of using AI tools with specialized coding LLMs to develop software, is making waves, what are the implications for security teams? How can this new way of developing applications be made secure? Or have the horses already left the stable? Segment Resources: https://www.backslash.security/press-releases/backslash-security-reveals-in-new-research-that-gpt-4-1-other-popular-llms-generate-insecure-code-unless-explicitly-prompted https://www.backslash.security/blog/vibe-securing-4-1-pillars-of-appsec-for-vibe-coding This segment is sponsored by Backslash. Visit https://securityweekly.com/backslashrsac to learn more about them! The rise of AI has largely mirrored the early days of open source software. With rapid adoption amongst developers who are trying to do more with less time, unmanaged open source AI presents serious risks to organizations. Brian Fox, CTO & Co-founder of Sonatype, will dive into the risks associated with open source AI and best practices to secure it. Segment Resources: https://www.sonatype.com/solutions/open-source-ai https://www.sonatype.com/blog/beyond-open-vs.-closed-understanding-the-spectrum-of-ai-transparency https://www.sonatype.com/resources/whitepapers/modern-development-in-ai-era This segment is sponsored by Sonatype. Visit https://securityweekly.com/sonatypersac to learn more about Sonatype's AI SCA solutions! The surge in AI agents is creating a vast new cyber attack surface with Non-Human Identities (NHIs) becoming a prime target. This segment will explore how SandboxAQ's AQtive Guard Discover platform addresses this challenge by providing real-time vulnerability detection and mitigation for NHIs and cryptographic assets. We'll discuss the platform's AI-driven approach to inventory, threat detection, and automated remediation, and its crucial role in helping enterprises secure their AI-driven future. To take control of your NHI security and proactively address the escalating threats posed by AI agents, visit https://securityweekly.com/sandboxaqrsac to schedule an early deployment and risk assessment. Show Notes: https://securityweekly.com/asw-332

In the news, Coinbase deals with bribes and insider threat, the NCSC notes the cross-cutting problem of incentivizing secure design, we cover some research that notes the multitude of definitions for secure design, and discuss the new Cybersecurity Skills Framework from the OpenSSF and Linux Foundation. Then we share two more sponsored interviews from this year's RSAC Conference. With more types of identities, machines, and agents trying to access increasingly critical data and resources, across larger numbers of devices, organizations will be faced with managing this added complexity and identity sprawl. Now more than ever, organizations need to make sure security is not an afterthought, implementing comprehensive solutions for securing, managing, and governing both non-human and human identities across ecosystems at scale. This segment is sponsored by Okta. Visit https://securityweekly.com/oktarsac to learn more about them! At Mend.io, we believe that securing AI-powered applications requires more than just scanning for vulnerabilities in AI-generated code—it demands a comprehensive, enterprise-level strategy. While many AppSec vendors offer limited, point-in-time solutions focused solely on AI code, Mend.io takes a broader and more integrated approach. Our platform is designed to secure not just the code, but the full spectrum of AI components embedded within modern applications. By leveraging existing risk management strategies, processes, and tools, we uncover the unique risks that AI introduces—without forcing organizations to reinvent their workflows. Mend.io’s solution ensures that AI security is embedded into the software development lifecycle, enabling teams to assess and mitigate risks proactively and at scale. Unlike isolated AI security startups, Mend.io delivers a single, unified platform that secures an organization’s entire codebase—including its AI-driven elements. This approach maximizes efficiency, minimizes disruption, and empowers enterprises to embrace AI innovation with confidence and control. This segment is sponsored by Mend.io. Visit https://securityweekly.com/mendrsac to book a live demo! Show Notes: https://securityweekly.com/asw-331

Developers are relying on LLMs as coding assistants, so where are the LLM assistants for appsec? The principles behind secure code reviews don't really change based on who write the code, whether human or AI. But more code means more reasons for appsec to scale its practices and figure out how to establish trust in code, packages, and designs. Rey Bango shares his experience with secure code reviews and where developer education fits in among the adoption of LLMs. As businesses rapidly embrace SaaS and AI-powered applications at an unprecedented rate, many small-to-medium sized businesses (SMBs) struggle to keep up due to complex tech stacks and limited visibility into the skyrocketing app sprawl. These modern challenges demand a smarter, more streamlined approach to identity and access management. Learn how LastPass is reimagining access control through “Secure Access Experiences” - starting with the introduction of SaaS Monitoring capabilities designed to bring clarity to even the most chaotic environments. Secure Access Experiences - https://www.lastpass.com/solutions/secure-access This segment is sponsored by LastPass. Visit https://securityweekly.com/lastpassrsac to learn more about them! Cloud Application Detection and Response (CADR) has burst onto the scene as one of the hottest categories in security, with numerous vendors touting a variety of capabilities and making promises on how bringing detection and response to the application-level will be a game changer. In this segment, Gal Elbaz, co-founder and CTO of Oligo Security, will dive into what CADR is, who it helps, and what the future will look like for this game changing technology. Segment Resources - https://www.oligo.security/company/whyoligo To see Oligo in action, please visit https://securityweekly.com/oligorsac Show Notes: https://securityweekly.com/asw-330