←
Evan Sultanik3,055 posts
Evan Sultanik avatar

Evan Sultanik

@ESultanik

Ph.D. computer security researcher @TrailOfBits. Editor of and frequent contributor to #pocorgtfo. My CV is a PDF that’s also an NES ROM sultanik.com/nesresume/

◷ Joined December 2008▧ 236 Photos and videos
Evan Sultanik avatar
Evan Sultanik@ESultanik·15 Jun 2020

After 6 months and over 5k new lines of 6502 assembly, the Kaizo-style platforming section of the NES game in my résumé is finally done! Yes, among other things, the PDF of my résumé is also an NES ROM. You can download it here for your emulating pleasure: sultanik.com/files/ESultanik…

Post attachmentPost attachmentPost attachment
Evan Sultanik avatar
Evan Sultanik@ESultanik·Nov 10

Morals of the story: - Follow @dguido's advice from the quoted thread to harden VS Code and GitHub; - Don't blindly pull code from a PR, even if you've previously reviewed it; and - OpenAI's Dot is surprisingly proactive!

Evan Sultanik avatar
Evan Sultanik@ESultanik·Nov 10

The scale was large enough that GitHub appears to have banned the contributor's account (it now 404s). This deleted all of its PR branches, causing all of its open PRs to 404. The side effect of this is that all of its original, benign, pre-force-push PRs instantly disappeared.

Evan Sultanik avatar
Evan Sultanik@ESultanik·Nov 10

This was a part of a larger campaign: The compromised account made similar force-pushes to many other repos, including CPython.

Evan Sultanik avatar
Evan Sultanik@ESultanik·Nov 10

My OpenAI Dot alerted me to the suspicious activity a few hours later and warned me not to pull the changes. It's unclear what prompted it to check; I never explicitly asked it to monitor my GitHub repos! @dguido's own Dot alerted him a couple hours later.

Evan Sultanik avatar
Evan Sultanik@ESultanik·Nov 10

After multiple developers had locally pulled the PR branch for testing/review, the external contributor force-pushed a new history for the PR that contained malicious VS Code scripts. Any dev who `git pull`ed afterward with a default VS Code config would have been compromised!

Evan Sultanik avatar
Evan Sultanik@ESultanik·Nov 10

Some context on this attempted supply chain attack on Graphtage, a tool I maintain… TL;DR: A trusted external contributor submitted a PR fixing an issue. CI was green, but we requested changes. Then the contributor's GitHub account was presumably compromised… 🧵

My Dot found malicious code in an open PR on a @trailofbits Github repo. I didn't ask it to hunt for malware! It noticed a commit, inspected what changed, and warned me about an attack all on its own.

Post attachment
Trail of Bits avatar
⟳ Evan Sultanik retweeted
Trail of Bits@trailofbits·Jun 5

We beat Google's quantum circuit again, and we didn't have to forge a proof this time. Today we're releasing trailmix, a toolkit for quantum "kickmix" circuits. It includes 5 new circuits we built for elliptic curve addition, the hardest part of Shor's algorithm.

Post attachment
Trail of Bits avatar
⟳ Evan Sultanik retweeted
Trail of Bits@trailofbits·25 Aug 2025

Solving the Traveling Salesman Problem for NYC's 474-station subway network, obviously! @ESultanik used Christofides algorithm to find a 20h 42min route through all 474 stations, which would beat the world record by 45 minutes. blog.trailofbits.com/2025/08…

Suha avatar
⟳ Evan Sultanik retweeted
Suha@suhackerr·21 Aug 2025

New post and tool! Attackers can break production AI systems by using image scaling to hide multi-modal prompt injections from users. 🧵for more info on what broke, how this works, and our new tool to try this out yourself

We hacked Gemini CLI, Vertex AI, Assistant, and other AI systems by embedding prompts into images that are not visible to users.

Post attachment
Trail of Bits avatar
⟳ Evan Sultanik retweeted
Trail of Bits@trailofbits·5 Feb 2025

Our new whitepaper covers secure-by-design steps that CEXes can take to keep users' accounts (and funds) safe from account takeover (ATO) in 2025. (Read more 👇)

Post attachment