←
Johnny Wang175 posts
Johnny Wang avatar

Johnny Wang

@wang_johnny

security @hinokisecurity. ex @google

◷ Joined August 2011▧ 26 Photos and videos
Johnny Wang avatar
Johnny Wang@wang_johnny·Oct 21

86% on CyberGym. On par with OAI Daybreak model (GPT-5.6-Sol), but with 0% cyber refusals. Introducing Hinoki Offensive One, the only open-access LLM designed for offensive security. Pentesting but getting refusals? Don't wait on PHP2Shell and PwnEscape. We can help ⬇️

Johnny Wang avatar
Johnny Wang@wang_johnny·3h

Thanks @AnthropicAI! Impressive launch with Claude Code. Have you considered looking for security vulnerabilities with @hinokisecurity? Our applied security engineers are happy to dig into your Claude Codee. Worth 5 minutes?

Post attachment
Johnny Wang avatar
Johnny Wang@wang_johnny·Nov 10

Ant's S-1: If our exclusionary and anticompetitive coalition known as Glasswing is exposed, we might lose our ability to commercialize Mythos, and our regulatory capture from safety-driven FUD would fall apart. @DarioAmodei Have you considered hiding this from your S-1?

Post attachment
Joel de la Garza avatar
⟳ Johnny Wang retweeted
Joel de la Garza@YODELAGARZA·Nov 9

If you haven't already - go run ARTEX against your infra. @hinokisecurity has made it incredibly easy and cheap to spin up for testing.

Try ARTEX before attacker runs it on your domain: @hinokisecurity now has ARTEX support! - Ask it to divide up your attack surfaces into sections ➗ - Use parallel agents to probe for vulns faster 🕵️ - Loop it for continuous monitoring ♾️ Try at hinoki.security

Post attachment
Johnny Wang avatar
Johnny Wang@wang_johnny·Nov 9

Try ARTEX before attacker runs it on your domain: @hinokisecurity now has ARTEX support! - Ask it to divide up your attack surfaces into sections ➗ - Use parallel agents to probe for vulns faster 🕵️ - Loop it for continuous monitoring ♾️ Try at hinoki.security

Post attachment
Sean Heelan avatar
⟳ Johnny Wang retweeted
Sean Heelan@seanhn·Oct 30

haha, so apparently OpenAI are about to start training on your Daybreak Blue sessions, with no way to opt out. Literally stealing people's private work, as they do it, is cartoonishly evil. It also will presumably kill Daybreak usage by security researchers. Or at least, it should. The thing OAI have had to swear they are not doing to math researchers, they are now explicitly about to start doing to security researchers. Yet another reminder of the importance of open models.

Johnny Wang avatar
Johnny Wang@wang_johnny·Oct 27

There are a lot of Brotli exploits out there. Be careful around whether they work on real environments. Getting around ASLR is very much not easy (even with the best models!), and brute-forcing approach that works on your quiet localhost might not work on prod.

Damián🦞 avatar
⟳ Johnny Wang retweeted
Damián🦞@fagamericano·Oct 26

Ran a Maestro / @Openclaw enterprise security assessment using an abliterated GLM 5.3 model thanks to @hinokisecurity and it gave us a B- on our architecture and deployment :) Nothing critical but we do have to enable better egress rules, it thankfully wasn’t able to steal a cred or token so I call that a success 🤣☺️