The Webpage Has Instructions. The Agent Has Your Credentials(openguard.sh)

22 points by everlier 7 months ago | 10 comments

Launch HN: Alter (YC S25) – Access control for multi-user agents(alterauth.com)

60 points by simfinii 5 days ago | 44 comments

Landing page for Alter, a YC S25 startup offering a proxy that manages credentials, policies, and audit logs for AI agents acting on behalf of multiple users.

Applications where agents are first-class citizens(every.to)

15 points by chrisjj 8 months ago | 12 comments

We found 24 Android vulnerabilities using our open source AI security agent(github.blog)

6 points by fourfire 27 days ago | 2 comments

GitHub Blog post describing how GitHub's open source AI security agent and its taskflows found 24 Android app vulnerabilities, and how to run it yourself.

Flock Hardcoded the Password for America's Surveillance Infrastructure 53 Times(nexanet.ai)

456 points by fuck_flock 9 months ago | 147 comments

Nexanet threat intel blog post reporting that Flock Safety hardcoded a password 53 times across its surveillance infrastructure software.

Claude Fable 5: mid-tier results on coding tasks(endorlabs.com)

373 points by bugvader 4 months ago | 217 comments

Endor Labs blog post benchmarking Claude Fable 5 on 200 real-world coding tasks for its Agent Security League, reporting middling functional and security results.

AI will make formal verification go mainstream(martin.kleppmann.com)

782 points by evankhoury 10 months ago | 394 comments

Martin Kleppmann blog post predicting that AI will bring formal verification with proof assistants like Rocq, Isabelle and Lean into mainstream software engineering.

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.

ChatGPT Enterprise(openai.com)

808 points by davidbarker 3 years ago | 452 comments

OpenAI blog post announcing ChatGPT Enterprise, a business version of ChatGPT with enterprise-grade security, privacy, and enhanced capabilities.

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.

Security auditing in the age of (good enough) AI(blog.trailofbits.com)

88 points by aray07 1 month ago | 13 comments

Trail of Bits blog post on using AI agents to build custom tooling (LSP server, decompiler, static analyzer) and formal models for auditing the Miden zero-knowledge VM.

Open source security at Astral(astral.sh)

344 points by vinhnx 6 months ago | 95 comments

Astral blog post by William Woodruff describing the techniques Astral uses to secure its open source tools and their build, test, and release processes against supply chain attacks.