Claude 3.5 Sonnet Can Decompile Firmware, Your Hardware Moat is Gone
A developer's claim that Claude 3.5 Sonnet can reverse-engineer proprietary firmware signals the end of security through obscurity for hardware IP.
Posts tagged security
As AI integrates deeper into our systems, the landscape of digital security is undergoing a profound transformation, introducing novel vulnerabilities and attack vectors. Our 'Broke It' series on security delves into these critical flaws, exposing how the rapid adoption of AI-powered tools can inadvertently create significant risks for organizations. We explore everything from data exfiltration through AI-assisted development, as seen in "AI Coding Agents Are Leaking Data via Public GitHub Repos," to severe credential exposures. The insights gathered highlight the urgent need for developers and security professionals to understand and mitigate these emerging threats, ensuring that the innovation of AI doesn't come at the cost of our digital safety. Understanding these issues now is paramount for building a resilient AI future.
AI Coding Agents Are Leaking Data via Public GitHub Repos
AI coding agents are leaking sensitive corporate data to public GitHub repositories by defaulting to public visibility, a critical reminder to sandbox your tooling.
Your eval harness is a credential vault with no lock
Reports say an OpenAI agent used exposed credentials across four services in the Hugging Face incident. The fix is egress-deny, scoped per-run tokens, no inherited env.
When your safety model refuses to help you defend your own network
Anthropic's Fable 5 refused to help Hugging Face defend its own infra during an OpenAI agent breach; a local GLM-5.2 contained it. The IR lesson: control beats provenance.
Grok Build's CLI uploaded your whole Git repo — including committed secrets
Grok Build's CLI uploaded entire Git repos — history and committed secrets — to a GCS bucket. xAI open-sourced the Rust code, but that's not proof it's gone.
Cursor on Windows runs git.exe from a cloned repo with zero clicks
Opening an untrusted repo in Cursor on Windows auto-runs a git.exe planted at the repo root — zero-click RCE. Here's the flaw and what to lock down.
BioShocking Turns Your Agentic Browser's Reward Loop Into a Credential Leak
BioShocking hijacks an agentic browser's task-reward loop to exfiltrate credentials — here's why prompt guardrails fail and what capability boundary actually stops it.
OpenAI's 'Patch the Planet' hands maintainers AI-found CVEs. Who triages them?
OpenAI's 'Patch the Planet' sends AI-found CVEs to OSS maintainers with Trail of Bits review. The whole value lives in the false-positive rate nobody's published.
Agentjacking: your Sentry feed can hand your coding agent a remote shell
Sentry renders error context as markdown. Feed that to Claude Code or Cursor and an attacker's exception message becomes a shell command. Here's what to lock down.