←
FAMASoon4,317 posts
FAMASoon avatar

FAMASoon

@FAMASoon

Security research blog: x64.moe GitHub: github.com/famasoon

◷ Joined April 2016▧ 187 Photos and videos
Nathan avatar
⟳ FAMASoon retweeted
Nathan@calysteon·Nov 15

These CVEs grew out of a breakthrough in how I use local language models: making them more capable without making them bigger. The research that made this possible is now public, with the paper, code and experimental results: calysteon.github.io/research…

CVE-2026-65373, CVE-2026-112511, CVE-2026-84496, CVE-2026-113946, CVE-2026-113941, CVE-2026-84498, CVE-2026-84590, CVE-2026-112482, CVE-2026-112457, CVE-2026-112443, CVE-2026-112708, CVE-2026-112461, CVE-2026-112417, CVE-2026-112481, CVE-2026-112451, CVE-2026-112622, CVE-2026-112529, CVE-2026-65384, CVE-2026-112526, CVE-2026-112620, CVE-2026-112671, CVE-2026-112519, CVE-2026-112414, CVE-2026-112604, CVE-2026-112546, CVE-2026-112413

Post attachment
☩MalwareMustDie avatar
⟳ FAMASoon retweeted
☩MalwareMustDie@malwaremustdie·Nov 13

Just to put things right: 1/ Many read my #ARTEX signature release as "an IOC list". It isn't. The #IOCs expire within a year by design. The real release is the workflow: how to turn the next ARTEX-style attack into detections as fast as the attacker moves. 2/ ARTEX is an #LLM harness. Swap the model, prompts or infra and every IOC changes. What changes much less is how an agent behaves: machine-speed recon chains, ~1/s #PasswordSprays with rotating UAs, self-correcting API retries, systematic record enumeration. 3/ So the pack is behavior-first. Those patterns live in Sigma correlations, Sentinel KQL and Zeek SumStats on the upper network layers, not in packet filters. IPs are clearly labelled T3/T4 context for retro-hunting what already happened. 4/ The #playbook is the core: verify the source, classify how specific an indicator really is, assign a tier, route it to the right format, expire it. We applied it to four campaigns with an LLM assisting and humans gating every verification step. Next trail, same loop. 5/ That's the big picture: attackers use LLMs to compress weeks into hours. Defenders can use LLMs to compress "new report → tested detection" into hours too, as long as humans keep the verification gates. We must defend atacker's powered LLM/#AI with AI. 6/ The Zeek section shows how to carry the same behavioral logic into network sensors you already run. Playbook, rules and tests: [link]. Break them, send FP/TP reports. That's how this grows.

#BLOG: #MMD-072-2026 - #ARTEX detection signature pack - from #IOC w/#playbook : blog.malwaremustdie.org/2026… Artex detection signature pack's repo: github.com/unixfreaxjp/malwa… Read the playbook carefully, asses & test before apply, you can use my python scripts in /tools dir and see /tests result in github repo beforehand. Posted to inspire #block-ing this threat for defenders in FIRST(dot)oeg or xCERTs and xSIRTs. Hope helps. - @unixfreaxjp Nov,2026

Post attachment
semiexp🍆☄️@技術書典21-か13 avatar
⟳ FAMASoon retweeted
semiexp🍆☄️@技術書典21-か13@semiexp·Nov 13

#技術書典 21 で SAT ソルバーの解説書を出します! SAT ソルバー「Glucose」のソースコードを細部まで読みながら、SAT ソルバーのアルゴリズムを詳しく解説する本になっています。

Post attachmentPost attachmentPost attachmentPost attachment
mizutani avatar
⟳ FAMASoon retweeted
mizutani@m_mizutani·Nov 7

来週のこの登壇、「専用の監視基盤とかなくてもお手軽にこういうところから始めてみるのはどうでしょう?」っていう話をするつもりだったのに、「いや、もう無理してでも監視基盤を用意してどうにかしないとダメじゃない…?」みたいな情勢に変わってしまい頭を抱えている tsumugi.io/events/btconjp-in…

Post attachment
MigawariIV avatar
⟳ FAMASoon retweeted
MigawariIV@strinsert1Na·Oct 31

本当に最後にこれだけ呟いて、あとはにゃーんする アトリビューションはフォレンジック的な側面だけで達成できるものではなく、物理や人的な介入も必須な長期戦。 公開情報のピボッティングは有益だが、高々2~3の情報の連鎖だけで plausible deniability を乗り越えることはかなり難しい。本気で社会のためを思うなら、捜査関係者に情報提供をしながら、複数の捜査機関と協力しながら裏の舞台でやって異なるベクトルからの証拠を固めるべき。 少なくともアトリビューションをSNSで公開しながらやっても捜査関係者の邪魔をしているだけで、得をするのは公開してインプレションを稼いだ人くらい。この構図はかつての私人逮捕系の流れと同じだし、賛同している人の気持ちも理解できない。

なかざん avatar
⟳ FAMASoon retweeted
なかざん@Nkzn·Oct 31

#KomeKaigi 2026の招待講演でお話しした「視野はAIに、視座は人に」の資料です。 普段から気にしてる人にとってはあたりまえ体操かもしれないし、迷ってる人には何かの道しるべになるかもしれない。 speakerdeck.com/nkzn/ai-for-…

Alex Neff avatar
⟳ FAMASoon retweeted
Alex Neff@al3x_n3ff·Oct 28

Bypassing SMB signing enforcement🔥 @aas_s3curity found a re-authentication bug in SMB. This allows you to simply relay an authentication to a host (e.g., a DC) even if signing is enforced! His NetExec module to test if a host is still vulnerable is now merged into main🚀 1/3🧵

Post attachment