Bishop Fox offensive security researchers and hackers take an unfiltered look at the latest cybersecurity headlines and give you a straight take. Do you actually need to care, or is it more of the same problems we’ve been seeing for years?
This episode covers a ChatGPT flaw that let one link spin up a fully authorized rogue AI agent, GitHub's public bug bounty cuts, Origin Energy's second breach in months, and 1,100+ AI lab employees asking Washington to help pace AI development, plus a sit-down with Bishop Fox's Dan Petro on catching cheaters in Super Smash Bros. Melee.
This episode breaks down OpenAI's own AI agent hacking into Hugging Face without human direction, an Android attack that hides commands in invisible screen text to hijack AI phone agents, and the White House's new Gold Eagle vulnerability clearinghouse, plus a sit-down with Bishop Fox's Emilio Gallegos on snowpick, his new open-source ServiceNow exposure scanner.
This special Red Team episode goes inside with the Bishop Fox Red Team, talking through what attribution actually looks like when device-level telemetry breaks anonymity, what young attackers with real skills but zero tradecraft have in common, and how one credential reset becomes the master key to everything downstream.
A three-year-old router backdoor makes fresh headlines, a phishing kit called EvilTokens hides until a victim's browser builds it, and CISA turns Anthropic's Mythos model on the government's own code.
This episode breaks down a Cisco flaw exploited within 24 hours of disclosure, a credential-harvesting campaign that hit 430,000+ FortiGate firewalls, and what Fable's restricted return and China's Tulongfeng mean for controlling offensive AI. Plus, a sit-down with Bishop Fox's Shad Malloy on building the first open-source Sparkplug B fuzzer for ICS environments.
This episode looks at how trusted systems become attacker pathways, from FIFA’s identity and authorization breakdown to FFmpeg’s downstream exposure, Klue’s abused OAuth access, and residential proxy networks that make malicious traffic look ordinary.
This episode explores what happens when the systems people trust quietly extend into domains they never agreed to. A Pokémon GO AR dataset trained a Visual Positioning System now adjacent to military drone navigation. A ServiceNow authentication flaw handed attackers read access to the operational core of enterprise IT. And the US government pulled two frontier AI models off the market over a jailbreak, with no established framework to bring them back.
This episode explores how the attack surface keeps expanding at every layer — from a single inverted kernel character enabling unauthenticated root, to AI assistants weaponized as system-wide IPC through notification injection, a Cisco SD-WAN zero-day giving attackers control of enterprise routing fabric, and the week's unavoidable elephant: whether Claude Fable V's guardrails actually hold.
This episode explores how attackers live in the gap between what a system can verify and what it settles for, from forged GlobalProtect VPN sessions to an autonomous AI worm, a social-engineered Meta support bot, voice-phished Salesforce access, and hotel reservation hijacking.
This special red team episode goes inside with the Bishop Fox Red Team — exploring how AI accelerates custom payload evasion and social engineering at scale, what a chained network-to-physical breach looks like in practice, and why satellites and gas pumps are reachable from the public internet right now.
This episode explores how attackers exploit infrastructure that became load-bearing before anyone secured it from a malicious VS Code extension that compromised thousands of GitHub repositories and an actively exploited Exchange zero-day, to Cisco SD-WAN auth bypasses, AI chaining low-severity bugs into real attack paths, and AWS GovCloud credentials left exposed in a public repo.
This episode explores how modern development's trust assumptions keep failing in attackers' favor, from the first confirmed AI-written zero-day to a coordinated supply chain attack poisoning 518 million download paths, developer credential harvesting via rootkit, AWS SES abuse for phishing at scale, and thousands of vibe-coded apps leaking sensitive data in the open web.
This episode explores how every layer of the stack has become an attack surface — from a privilege-escalating Linux kernel flaw and a GitHub infrastructure RCE to a poisoned RubyGems supply chain, a trojanized vendor installer, and a ransomware hit on centralized education infrastructure.
This episode explores how access is being created, scaled, and kept with less friction, from a critical cPanel authentication bypass to AI-generated vulnerable code, AI-assisted attacks, persistent footholds in trusted systems, and stealthier data exfiltration.
This episode explores how access control is breaking down across AI systems, consumer apps, and vulnerability management, from leaked AI tooling and bypassed EU verification apps to actively exploited Windows zero-days and growing strain on the NVD.
This week’s episode is about attackers working through what’s already trusted. Not broken. Not bypassed. Trusted.
In this special episode, we break down Anthropic’s Project Glasswing announcement and what it signals for the future of cybersecurity.
In this Initial Access podcast episode, we examine how trust, speed, and automation are reshaping initial access across software supply chains, network infrastructure, and AI systems.
In this Initial Access podcast episode, we examine how attackers are inheriting access through trusted systems, default permissions, and unpatchable infrastructure.
In this Initial Access podcast episode, we examine how attackers are turning normal workflows and trusted systems into reliable paths for initial access as exploitation timelines continue to shrink.
In this Initial Access podcast episode, the team looks at several recent examples of that compression in action, from a supply chain compromise that led to AWS admin access, to malware spreading through GitHub, npm, and VS Code, to ClickFix lures that convince technical users to run malicious commands themselves.
In this Initial Access podcast episode, we examine how attackers are gaining initial access through social engineering, identity abuse, and vulnerable edge infrastructure.
In this Initial Access podcast episode, we cover AI coding agents operating inside developer environments, automated attack platforms accelerating exploitation cycles, long-lived connected devices exposing unexpected telemetry risks, and why identity systems remain the primary entry point for attackers.
In this Initial Access podcast episode, we cover autonomous vulnerability discovery, AI agents that ignore instructions, and why models are becoming strategic national assets.
In this Initial Access podcast episode, we cover SSO phishing, patching failures, exposed APIs, and zombie infrastructure remind us that basic security hygiene still decides the outcome.
In this Initial Access podcast episode, we cover prompt injection, a hijacked Outlook add-in, commoditized mobile spyware, AI executive deepfake scams, IT-to-OT pivoting, and nation-state use of commercial LLMs to accelerate exploitation.
In this Initial Access podcast episode, we cover the rollback of federal software security guidance, insider-driven access risks, ongoing state-sponsored espionage, and the security implications of giving AI tools deep control over infrastructure.
This week, we took a real look at the latest security headlines and have a straight take on them. The goal is simple: do you actually need to care about this, or is it just another variation of the same fundamental security problems we’ve been dealing with for years?