NemoClaw – Nvidia's upcoming open-source AI agent platform(nemoclaw.bot)

13 points by umangsehgal93 7 months ago | 2 comments

A heap overflow and SSO misconfiguration to compromise OpenAI internal repos(hacktron.ai)

456 points by Handy-Man 1 month ago | 192 comments

Hacktron research blog post describing how a heap overflow and an SSO misconfiguration were chained to compromise OpenAI internal repositories.

GitHub Copilot CLI downloads and executes malware(promptarmor.com)

43 points by sarelta 8 months ago | 12 comments

Nostr – Decentralized social network(nostr.com)

124 points by throwaway689236 3 years ago | 123 comments

OpenClaw Exposure Watchboard(openclaw.allegro.earth)

36 points by fanweixiao 7 months ago | 16 comments

All your OpenCodes belong to us(johncodes.com)

33 points by jpmcb 9 months ago | 7 comments

How we hacked McKinsey's AI platform(codewall.ai)

420 points by mycroft_4221 7 months ago | 168 comments

CodeWall blog post describing how its autonomous offensive agent found vulnerabilities in McKinsey's internal AI platform Lilli.

Disrupting the first reported AI-orchestrated cyber espionage campaign(anthropic.com)

341 points by koakuma-chan 11 months ago | 261 comments

Anthropic news post describing how it detected and disrupted a cyber espionage campaign largely orchestrated using its Claude AI model, with a linked full report.

Anthropic’s paper smells like bullshit(djnn.sh)

1070 points by vxvxvx 11 months ago | 300 comments

Blog post criticizing Anthropic's report on disrupting an AI-orchestrated cyber espionage campaign; the captured page shows a Cloudflare block screen.

Claude Cowork exfiltrates files(promptarmor.com)

827 points by takira 9 months ago | 366 comments

PromptArmor threat intelligence report showing Claude Cowork is vulnerable to file exfiltration via indirect prompt injection due to code execution isolation flaws.

AI models can provably hide arbitrary capabilities(arxiv.org)

252 points by sbulaev 13 days ago | 118 comments

An arXiv paper showing AI models can embed encrypted neural network backdoors that provably hide arbitrary capabilities from elicitation-based safety evaluations.

Devs say AI crawlers dominate traffic, forcing blocks on entire countries(arstechnica.com)

329 points by LinuxBender 1 year ago | 232 comments

Ars Technica article reporting that open source developers face overwhelming AI crawler traffic, leading some to block entire countries to keep FOSS sites online.