Breach Please

August 27, 2026. S0:E22. We go a little long discussing the OpenAI post-mortem on Hugging Face.

Thu, 27 Aug 2026 04:00:00 -0400

In today's episode of Breach Please, Jake and Jess talk through the OpenAI post-mortem on the Hugging Face hack. It's a doozy. There's little doubt in either of our minds that the facts as presented don't exonerate OpenAI - they make it look far worse. We cut through the hype and conclude that there's no way OpenAI was following even the most basic of security best practices, even accounting for hindsight bias. Jake closes by reading a passage from one of Sam Altman's tweets from December 2025, which made clear he knew the dangers of AI - the kind of thing that is likely to haunt him in the future.

https://openai.com/index/hugging-face-incident-and-the-road-ahead/

https://x.com/sama/status/2004939524216910323

S0E21: Alabama Comes for OpenAI, WebLogic Comes for Everyone

Wed, 26 Aug 2026 07:00:00 -0400

Fifteen state attorneys general, led by Alabama, just subpoenaed OpenAI over the Hugging Face breach, demanding the company preserve all evidence related to the intrusion. Jess and Jake break down what that legal hold actually requires, why "mark it ACP" doesn't make a Slack channel privileged, and why the discovery list's question about internal safety concerns might be the part that burns OpenAI hardest.

Then: a perfect-10 WebLogic vulnerability, patched back in January, just landed on CISA's Known Exploited Vulnerabilities list, eight months after active exploitation began. Jess and Jake talk through why "patch applied" isn't the finish line and what a post-patch threat hunt should actually look like.

In this episode: Fifteen states subpoena OpenAI over the Hugging Face hack and order evidence preservation What a legal hold actually requires (and what evidence spoliation means) Why marking a channel "ACP" doesn't make it attorney-client privileged The discovery request digging into internal safety concerns and model testing A perfect-10 WebLogic CVE, patched in January, added to CISA's KEV eight months later Why "patch applied" doesn't mean "threat hunt done"

S0:E20 Multi-Agent AI Systems Turn Into Turf Wars

Tue, 25 Aug 2026 07:00:00 -0400

AI just became a cybersecurity governance problem in public. Jess Hebenstreit and Jake Williams break down Anthropic’s multi-agent research, where AI agents with conflicting goals start negotiating, forcing, or trucing their way through “collaboration.”What does it mean when one model settles by force and another settles by truce?

Jess and Jake unpack what that means for real-world agent deployments, human-in-the-loop approvals, and the danger of letting “AI says so” become the final word. You’ll learn why multi-agent systems can spiral into turf wars, why governance gets harder once agents are involved, and how to build controls that still let humans say no.

S0:E19: Cloud Platform Vulnerabilities and Vendor Transparency

Mon, 24 Aug 2026 06:15:00 -0400

In this episode, Jake and Jess talk about vulnerabilities in cloud platforms and how vulnerabilities should be disclosed. We discussed this in the context of recent Microsoft vulnerabilities in Entra, Azure Arc, Exchange Online, and Managed Apache Casandra. All of these had vulnerabilities with minimal details from Microsoft. We generally agree that this isn't what transparency looks like. We also discuss a patch from April that is now actively exploited in the wild.

S0:E18. Student stops AI supply chain attack and prompt injection is still a thing...

Fri, 21 Aug 2026 07:00:00 -0400

In this episode of Breach Please, we talk about the details that emerged about the Anthropic agent that tried to social engineer a college student into committing a supply chain attack. The whole thing feels very xz-utils 'esque, only with an agent - and a malfunctioning one at that.

Then we cover some interesting reporting by Dan Goodin showing how Grok's guardrails can be bypassed through passing it encrypted data along with the instructios to decrypt it. We reinforce that prompt injection is a feature of LLMs, not a bug.

We then talk about informed consent of the risks and applicable guardrails based on a viewer note. We even tee up the question of what cyber insurance is likely to cover costs of AI agent activity. Finally, we opine on whether orgs should consider including the inevitable costs of incidents in AI deployment budgets.

Show links:

https://www.reuters.com/world/how-texas-student-blew-whistle-rogue-ai-hacking-attempt-2026-08-20/

https://arstechnica.com/security/2026/08/grok-exfiltrates-user-data-when-malicious-instructions-are-encrypted/

S0E17: Amazon vs Perplexity Reveals the First Big AI Agent Liability Test

Thu, 20 Aug 2026 07:00:00 -0400

AI agents are about to test the limits of who gets blamed when they act on your behalf. A new Ninth Circuit ruling in the Amazon vs. Perplexity fight could reshape how enterprises think about autonomous tools, non-human identities, and legal exposure - and the takeaway is more unsettling than most vendors will admit.

Jake Williams and Jess Hebenstreit break down what happened when Amazon pushed back against Perplexity’s AI browser, why the CFAA and California’s CDFAA matter here, and how the court’s reasoning could shift responsibility from the agent publisher to the organization that deploys it. They also dig into what this means for legal teams, risk registers, and why every AI agent needs its own identity instead of borrowing a human user’s.

You’ll also hear how this ruling fits into the bigger enterprise governance problem: who owns AI risk, who should be accountable when an agent crosses a line, and why security teams can’t be expected to carry every line-of-business risk themselves. Jake and Jess make the case for splitting cybersecurity risk from business risk, naming a real owner, and making legal counsel aware before your agent does something expensive.

Then the conversation pivots to another cautionary tale: a 3M expert witness who allegedly used ChatGPT to generate a report aimed at proving the company was 0% at fault in a gas detector explosion case.

That story opens up a bigger discussion about leading prompts, hallucinated conclusions, discoverability, and why AI-generated work product can become evidence against you. Perfect for security leaders, legal teams, incident responders, and anyone deploying AI agents in the enterprise.

S0:E16. OpenAI Highlights Security Basics and Hopes We Won't Notice

Wed, 19 Aug 2026 07:00:00 -0400

In this episode of Breach Please, Jake and Jess talk through OpenAI's recent blog on "pacing model development." The blog reads pretty poorly and really just covers security best practices, apparently hoping readers won't notice that OpenAI is just recommending basics it should have been doing all along (monitoring, sandboxing, etc.). One significant takeaway is that OpenAI has probably raised the bar for agentic logging, owing to the apparent requirement to analyze reasoning traces. We talk about why this is difficult, at least in part because these will contain sensitive data not appropriate for most SIEMs.

Show links:

https://openai.com/index/pacing-model-development-cyber-capabilities/

https://www.custody-framework.org/

S0E15 Faking out Famous Chollima and is in-car infotainment advertising a threat?

Tue, 18 Aug 2026 07:00:00 -0400

In this episode of Breach Please Jake and Jess talk about Famous Chollima getting faked out and tricked into revealing TTPs. We discuss our own experience with DPRK workers and share actionable tips for not hiring them or discovering them after they've been hired. Jake may or may not have gone on a rant about cryptocurrency more generally and how it facilitates lots of crime (inspired by talking about DPRK using it for sanctions evasion).

Then we talk about in-car advertising on infotainment systems. These won't be patched reliably. Do we need regulation here? Jake thinks it's a right to repair issue.

Jake's class Friday August 21, 2026: https://learning.antisyphontraining.com/courses/assessing-ai-security-model-context-protocol

Jake's class at Wild West Hackin' Fest in October: https://www.antisyphontraining.com/product/hands-on-ai-security-risk-assessment-with-jake-williams/

Show notes: https://any.run/cybersecurity-blog/lazarus-group-it-workers-investigation-part-two/

https://disconnect.blog/bmws-in-car-spider-man-ad-is-a-warning-sign/

S0:E14 - CTI and Geopolitical Events

Mon, 17 Aug 2026 07:00:00 -0400

In this episode of Breach Please, we talk about whether CTI analysts can use changes in dwell times, cyber targeting, etc. as a leading indicator of impending geopolitical events. We generally think there's to much noise to be generally effective (especially for Taiwan).

Then we talk about a Mac vuln that's being exploited in the wild, with a CVSS score of 9.8. Somehow it isn't in CISA's KEV, but hey, what's a little vulnerability enrichment delay between friends...

Show links:

https://www.tomshardware.com/tech-industry/cyber-security/macos-screen-sharing-flaw-exploited-to-root-macs-and-plant-monero-miners

S0E13 - We talk about private companies conducting cyber ops and another AI oopsie.

Fri, 14 Aug 2026 07:00:00 -0400

In today's episode of Breach Please, Jake and Jess talk about cyber companies conducting cyber ops against transnational-criminal organizations, per the new Presidential directive. We also talk about a bug in how major AI platform providers encrypted their chain of thought traces for their frontier models. The vulnerability, when combined with jailbreaking of less powerful models, allowed the traces (previously claimed to be trade secrets) to be extracted.

https://www.whitehouse.gov/presidential-actions/2026/08/expanding-capabilities-to-combat-transnational-cyber-enabled-crime/

https://cybersecuritynews.com/top-ai-models-apis-flaw-exposes-hidden-reasoning/

S0E12 - Vulnerable edge devices, Fortinet exploitation, and whether firewalls belong in zero trust

Thu, 13 Aug 2026 07:00:00 -0400

This episode tackles why internet-facing edge devices keep showing up in major incidents and why that creates more than just patching work. Jess Hebenstreit and Jake Williams also dig into the real operational cost of keeping vulnerable gear in place, from constant incident response to the challenge of proving a device is actually clean. We discuss Fortinet exploitation tied to Gunra ransomware, why OT and IT environments make edge patching especially hard, and whether firewalls should be considered part of a zero trust architecture. Along the way, Jess and Jake compare the realities of vendor complexity, flat networks, and the limits of threat hunting after an exploit.

S0E11 - Using AI for patching and inflight shenanigans.

Tue, 11 Aug 2026 07:00:00 -0400

In this episode of Breach Please, Jake and Jess talk about new research showing that AI isn't actually that good at generating patches, highlighting objective rates of failure. Turns out, frontier models only generate the correct patch without introducing issues in about 25% of cases. In almost 5% of cases, the AI introduces new vulnerabilities. We also talk about some shenanigans on a Delta flight home from DEFCON. We discuss how time is a flat circle and reminisce Chris Roberts' issues with in-flight hacking from 2015.

Show links:

https://1password.com/blog/why-ai-generated-patches-still-require-human-review https://www.defendersinitiative.com/p/reviewing-initial-research-on-using https://www.insideflyer.com/posts/hackers-spoofed-delta-wi-fi-on-las-vegas-flight/

S0E10 - Commonalities in AI lab escapes and How to govern smart devices.

Thu, 06 Aug 2026 07:00:00 -0400

In this episode of Breach Please, Jake and Jess talk about governing smart devices in the enterprise, pivoting off some outstanding research work that Andy Greenberg covered (and was a subject in himself). Then we talk about a common player whose name keeps popping up in these "AI lab escape" disclosures (and who we saw *again* with the Meta disclosure). We bat around some actionable advice for those thinking of outsourcing AI lab containment. Jess is stress testing the Breach Please stickers propping up our banner behind us. Throughout the episode, we see that while they are objectively great stickers, they are not great for this task.

Show links:

https://www.wired.com/story/hackers-stalked-me-by-hijacking-a-smartwatch-for-kids/ https://www.bleepingcomputer.com/news/security/meta-ai-model-hacked-a-company-during-misconfigured-cyber-test/ https://www.custody-framework.org/

S0E9 - AI social engineering & operational concerns with AI workflows.

Wed, 05 Aug 2026 07:00:00 -0400

In this episode, Jake and Jess talk about more issues with AI agents escaping containment, this time actively using social engineering to compromise victims. We pose ethical questions about whether you have a duty to investigate your logs for agent breakouts. Then, we discuss how Grokipedia can serve as a cautionary tale in AI workflows silently breaking down.

Show links: https://www.bleepingcomputer.com/news/security/openai-anthropic-ai-agents-targeted-real-people-and-systems-in-cyber-tests/ https://www.lawfaremedia.org/article/grokipedia-stopped-reviewing-edits-in-april.-it-didn-t-tell-anyone

S0E8 - Anthropic ruins Jake's best laid plans, BMC controller vulns, and SentinelOne vs HJ

Mon, 03 Aug 2026 07:00:00 -0400

In this episode of Breach Please, Jake and Jess talk about Anthropic announcing they have lost control of their agents too (not to be outdone by OpenAI). OpenAI is in good company though, since Anthropic ALSO failed to notice their agents breaking out of their sandbox environments. Jake mentions CUSTODY, a framework he'd been planning to present on later this year until these irresponsible agentic containment issues forced him to release it today for the good of the community. Then we talk about BMC controller vulnerabilities that have their roots in weaknesses so old they can not only vote, but legally buy alcohol in all 50 states. Finally, we wrap up with some notes about vendor shenanigans at DEF CON, brought to you by Sentinel One. Hey friends, if you're planning on introducing a new vulnerability class like line hijacking, talk to the conference and/or the guy running the event first? Also, just don't. In a world where you can be anything, don't be a Sentinel One...

S0E7 - HuggingFace Post Mortem Breakdown

Fri, 31 Jul 2026 07:00:00 -0400

Jake and Jess discuss the excellent post mortem Hugging Face released as a follow on to the OpenAI agent attack. We cover everything from should we call an Autonomous Agent a Threat Actor, to API Security, other lessons learned.

S0E6 - Hacker Summer Camp Survival Guide

Thu, 30 Jul 2026 07:00:00 -0400

In this episode, Jake and Jess share their tips for doing Hacker Summer Camp (Black Hat, DEF CON, B-Sides, and SO MANY side conferences) the right way. If you've never been, things can be overwhelming and it's easy to overdo it. Even if you're an old hat, there's probably learnings here - plus Jess shows off some of the Breach Please swag we'll have with us.

S0E5 - Probable alert fatigue endangers DHS collaborators and duress passwords.

Wed, 29 Jul 2026 07:00:00 -0400

In this episode, Jake and Jess talk DHS (twice). First, we talk about DHS dismissing real alarms as false positives in what Jake assesses was a probable watering hole attack nobody seems to be acknowledging. Then we talk about a #stopCopCity activist being charged for giving CBP a duress password that wiped his device. While I think we should all agree that's not a chargeable offense, are duress passwords appropriate for enterprise applications? Jake and Jess debate the use of these passwords for enterprise applications. https://www.nextgov.com/cybersecurity/2026/07/dhs-network-intrusion-was-twice-ruled-false-positive-breach-confirmed/414724/ https://boingboing.net/2026/07/25/grapheneos-duress-password-border-search.html

S0E4 - Google indexes AI chats (again) and dump of PoC exploits

Tue, 28 Jul 2026 07:00:00 -0400

In this episode of Breach Please, Jake and Jess talk about how the story (again) about AI chats being indexed by search engines isn't really an AI story. This is a data security problem and a misalignment with enterprise tooling. We then talk about a relatively new GitHub repo with lots of zero day exploits. We agree it's academically interesting, but doesn't change anything for how we do enterprise security.

S0E3 - OpenAI Lost Control of its Agents

Mon, 27 Jul 2026 07:00:00 -0400

In this episode, we talk a lot (too much, we went over on time) about the reports that OpenAI lost control of its agents and apparently only realized it after Hugging Face posted and someone said "yeah, that sounds a bit like us." Reuters article: https://www.reuters.com/business/its-ai-agent-spent-days-hacking-company-sources-say-openai-did-not-notice-week-2026-07-24/ Jake mentioned training. Here's an upcoming one-day seminar on MCP security: https://learning.antisyphontraining.com/courses/assessing-ai-security-model-context-protocol Here's Jake's two-day class covering AI security risk assessment: https://www.antisyphontraining.com/product/hands-on-ai-security-risk-assessment-with-jake-williams/

S0E2 SNow vs researchers and destructive cyberattacks.

Fri, 24 Jul 2026 07:00:00 -0400

In this episode, Jake and Jess talk about lessons we can take away from the alleged exploitation of ServiceNow (likely CVE-2026-6875) and the destructive attack on Romania's Land Registry Database. Join us as we discuss the stories, but more importantly what they mean for you and your security team.

S0E1 OpenAI hacked Hugging Face? Oh noes!

Thu, 23 Jul 2026 07:00:00 -0400

Jake and Jess talk through the story (and there's a LOT missing here) where OpenAI's agent allegedly hacked Hugging Face. There's so much meat missing in OpenAI's account of what happened that their PR people are clearly advocating we go vegetarian...

S0E0 - Is This Thing On?

Wed, 22 Jul 2026 07:00:00 -0400

In this episode, we discuss our mission statement. What are we even doing with Breach Please? What can you expect? Does the industry even need more talking heads? We think so - but only if they tell it like it is...