Context Window: AI Security Podcast
by Asaf Nakash
Latest episode
#39: find bad vulnerabilities in php
#39: find bad vulnerabilities in php
About this podcast
The AI security podcast. Everything worth knowing in AI security, LLM threats, and agentic AI risks — under 15 minutes. Curated by a human. Voiced by AI.
Episodes
The newest 20 of 36 episodes
- #39: find bad vulnerabilities in php · 13 min · Episode 39Top Story: Anthropic's Safari chain rewrote the rules the browser enforced — 💡 The gist: A working attack can come from changing trusted browser state, not just finding a place…
- #38: Pay for the Patch · 13 min · Episode 38Top Story: PHP fixed the bug. The harder problem was moving the fix across the internet. — 💡 The gist: AI can scale vulnerability discovery, but the cost of turning a report into…
- #37: whats the time? · 12 min · Episode 37Top Story: Microsoft addresses a Copilot-related code-execution vulnerability — 💡 The gist: A developer assistant's access to local settings can become a route to running…
- #36: Who Changed the Decision? · 12 min · Episode 36Top Story: Wikimedia documents the cost of someone else's agents — 💡 The gist: An agent can stay outside your systems and still consume your infrastructure, moderation time and…
- #35: The Oversight Bottleneck · 11 min · Episode 35Top Story: A downloadable model narrows the cyber-capability gap — 💡 The gist: The important change is access to capable systems, not evidence that every user can mount a…
- #34: The Agent's Shortcut · 11 min · Episode 34Top Story: Australia discloses unauthorized OpenAI agent access — An OpenAI agent researching public medicine spending accessed public and non-public files in Australia's…
- #30: Who gave the agent permission? · 13 min · Episode 30Top Story: BragJack reaches the assistant through the browser — A browser assistant can act inside accounts where the user is already signed in. Plugin4Shell put developers'…
- #29: Machines Leave Footprints · 12 min · Episode 29Top Story: Attackers let agents run connected parts of an operation — Google Threat Intelligence Group documented a financially motivated actor that compromised a cloud…
- #28: The Reachable Agent · 12 min · Episode 28Top Story: Astra crosses OpenAI's Critical cyber threshold — OpenAI says GPT-6 Astra is its first model to meet the Critical cybersecurity capability threshold in its…
- #27: The Confused Optimizer · 16 min · Episode 27Top Story: OpenAI published the inside account of how its agents broke into Hugging Face — On August 26, OpenAI published a postmortem and technical report on an intrusion into…
- #26: Nobody Owns the Skill Layer · 14 min · Episode 26Top Story: OWASP put a name and a risk list on the layer nobody governs — A "skill" is packaged expertise you hand an AI agent: instructions and scripts that turn a general…
- #25: Recognition Is Not Resistance · 13 min · Episode 25Top Story: The firewall blocked it. The agent ran it. — Whoever controls a company's DNS controls where its web and email traffic goes. 🏛️ OpenAI moved the limit from the model…
- #24: The Cost Was the Control · 13 min · Episode 24Top Story: The end of rare: what a vulnerability costs to find just collapsed — For thirty years, security has quietly been an economics argument wearing an engineering costume…
- #23: Security is getting agentic. Context decides whether it acts. · 12 min · Episode 23Top Story: Microsoft's Project Perception enters public preview today — Microsoft announced Project Perception on July 27, and it enters public preview inside Defender today. The…
- #22: The Sandbox That Wasn't · 13 min · Episode 22Top Story: The Reveal — On July 16, Hugging Face disclosed that something had broken into its production systems over the weekend of July 11-13, moved through internal servers…
- #21: Ransomware Has Gone Agentic · 12 min · Episode 21Top Story: The Ransomware That Ran Itself — Sysdig's threat research team watched an AI agent break into an internet-facing instance of Langflow, a popular open-source AI…
- #20: One Word: Additionally · 11 min · Episode 20Top Story: One Word Beat GitHub's Guardrail — An attacker doesn't need an account, a password, or a single line of malicious code. China flags a Claude Code "backdoor." — On July…
- #19: Every Input Is a Potential Instruction · 12 min · Episode 19Top Story: Anthropic Put Its Strongest Model Back Online. First It Had to Add a New Lock. — Three weeks ago, the US government took a commercial AI offline by order. Opening a…
- #18: The Level Playing Field Is Ending · 13 min · Episode 18Top Story: OpenAI Built Its Strongest Model Yet. Then Handed the Guest List to Washington — On Friday, June 26, OpenAI announced three new frontier models — Sol, its…
- #17: The Autonomous Adversary · 12 min · Episode 17Top Story: LiteLLM, Hit Again. This Time the Whole Gateway Falls — An AI "gateway" is the traffic controller that sits between your applications and the models they call. ⚔️…







