6 min read
A Record of Approval Is Not Approval
Darktrace rewrote the saved conversations of four coding agents and got them to attack a network. The lesson isn't about signatures. It's about where your agent's permission comes from.
Tagged: boundaries
8 posts tagged with “boundaries”
9 min read
Project Glasswing: AI Finds Zero-Days Faster Than Humans Can Patch Them
Anthropic's Project Glasswing deployed Claude Mythos Preview to autonomously discover thousands of zero-days with a 72.4% exploit success rate. Less than 1% of findings have been patched. The bottleneck is no longer discovery — it's everything that comes after.
12 min read
Vibe Coding's $1.5M Mistake
A penetration testing firm audited 15 applications built with AI coding assistants. They found 69 exploitable vulnerabilities, 6 critical. The estimated remediation cost: $1.5 million. Teams shipping AI-generated code need to focus on the security debt accumulating underneath.
9 min read
Guardrails Failed. Now What?
Static AI guardrails are failing in production. Langflow was exploited within 20 hours. Cline was compromised through a GitHub issue title. Here's what actually works instead.
12 min read
When AI Finds Bugs in Its Own Code
Claude found 22 Firefox vulnerabilities in two weeks. XBOW discovered a 9.8-rated Windows CVE without source code. But 50% of AI-generated code ships with exploitable flaws. AI is better at finding bugs than avoiding them — and that gap defines the next era of application security.
14 min read
The AI Agent Supply Chain Is Already Compromised
820 malicious packages. 30,000 exposed instances. Fortune 500 breaches. The AI agent ecosystem has a supply chain problem that traditional AppSec isn't built to catch.
8 min read
Your Agent's Real Attack Surface Isn't Its Prompt
Everyone optimizes the token window. Almost nobody manages the environment. Active context is what your agent thinks about. Latent context is what your agent can reach. The blast radius of a compromised agent is determined by the latter.
11 min read
Prompt Injection Goes Live: Three Proof Points That Change Everything
Indirect prompt injection has moved from theory to active exploitation. Unit 42 confirms in-the-wild attacks, PleaseFix hijacks AI agents through calendar invites, and a Claude Code CVE exposed 150,000 developers. Here is what security teams need to know.