RootSploit (@rootsploit) header
RootSploit (@rootsploit) profile picture

AI Red Team | Offensive Security Engineer @CrusoeAI | Ex @Amazon, @Apple | Bug Hunter | OSCP | Cobalt Core | Views are my own!!

Joined June 2020
Photo by RootSploit (@rootsploit)Photo by RootSploit (@rootsploit)Photo by RootSploit (@rootsploit)Photo by RootSploit (@rootsploit)Photo by RootSploit (@rootsploit)Photo by RootSploit (@rootsploit)Photo by RootSploit (@rootsploit)Photo by RootSploit (@rootsploit)Photo by RootSploit (@rootsploit)Photo by RootSploit (@rootsploit)
RootSploit retweeted
OpenRouter says a provider was hacked and served the wrong model for days. Would you have noticed? So @BugBunny_ai we fingerprinted all 50 DeepSeek V4.1 Flash endpoints on OpenRouter with 130k one-word answers. No model swaps. But 3 providers drift clearly from DeepSeek's own API. 🧵
11
7
4
107
13,084
RootSploit retweeted
Java4Shell: write once, RCE everywhere calif.io/java4shell
59
694
144
2,676
234,669
RootSploit retweeted
🚨👾First post → Awesome research. We compromised the latest version of OpenAI Codex. We prompt injected. We escaped a sandbox. We exfiltrated secrets… And ended up with control over what could be published to the official openai/codex npm package. Prompt injection → RCE → supply chain compromise. Full research by @0xMadder and @_Worty in the comments 👇
32
61
17
153
214,406
the plot thickens! 👀 turns out you can hide a secret message inside the special tag that Claude will initially deny awareness of, but that's a lie! Claudeception? 😵‍💫
interesting little tidbit about Claude’s reasoning effort notation! in the web app, this tag gets attached to the end of Claude’s prompt on every turn: <antml:reasoning_effort>8</antml:reasoning_effort> low = 8, medium = 10, high = 30, extra = 80, max = “max”
71
136
16
3,106
134,649
RootSploit retweeted
Qwen3.8 flash at 150–177 tok/s gen and ~1,900 tok/s prefill on 2x 3090s + 96gb ddr5 bone engine, my own inference engine. 256k context, single stream or 4x concurrent (50ts a stream, up to 256k) vs strata engine on the same rig: 120-140gen, 1k pp → 150-170gen, 2k pp iq3_xxs quant, so not full precision, will optimize q4 once qwen 4 drops. Model is leftover from when i had less ram.
3
1
15
785
🚨 CRITICAL - Unauthenticated RCE in React Server Components (GHSA-xqc2-xf4c-x3h3) A critical vulnerability in React Server Components packages (react-server-dom-webpack, react-server-dom-turbopack, react-server-dom-parcel) allows unauthenticated remote attackers to execute arbitrary code via crafted RSC Flight payload responses transporting multiple typed arrays and attacker-controlled values. 👉 Affected: Versions 19.0.0–19.0.8, 19.1.0–19.1.9, 19.2.0–19.2.8, 19.3.0 | Upgrade immediately to 19.0.9, 19.1.10, 19.2.9, or 19.3.1 (or Next.js 16.4.1 / 15.5.28) github.com/react/react/secur…
50
276
115
1,691
473,645
RootSploit retweeted
How to use LLMs for vulnerability research (@ZephrFish) MCP: blog.zsec.uk/bullyingllms/ Harness: blog.zsec.uk/harnessing-harn… #infosec #llm
15
166
1
1,270
61,678
RootSploit retweeted
Lockdown Mode helps, until it doesn't.
259
820
237
9,529
1,513,800
RootSploit retweeted
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵 Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-sp… github.com/RikiyaT/lm_und… huggingface.co/Qwen/Qwen3.5-0… huggingface.co/RikiyaT/Qwen3.…
2,368
7,253
5,757
55,904
10,098,645
RootSploit retweeted
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected. Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions. Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent. But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models? We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem. Read the full report: cohere.com/blog/concerning-b…
973
1,282
1,522
11,087
3,107,139
RootSploit retweeted
Published our high fidelity scanner for node2shell: github.com/assetnote/node2sh… Detection logic was written by @hash_kitten late last week. Open sourcing it to do public good in detecting this issue black box.
6
80
2
381
20,650
WOW did we really just get a Mythos-level open-weight model?? in OCTOBER?! 🤯 oh—hi there, past pliny! spot-on predictin, well struck fren! 🍻 and right on time 🎃👻 dunno how you do it! *takes a deep breath* everything IS going to be fine, isn’t it? thanks for the reminder 🤗 onwards! 🫡 ⊰-•-•✧•-•-⦑/L\O/V\E/,\P/L\I/N\Y/⦒-•-•✧•-•-⊱
Replying to @elder_plinius
if i were a gambler, i'd put my money on Halloween. can't wait to quote tweet this. hi frens from the future! take a deep breath! everything is going to be fine! ⊰•-•✧•-•-⦑/L\O/V\E/,\P/L\I/N\Y/⦒-•-•✧•-•⊱
86
143
21
2,605
104,059
RootSploit retweeted
Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all frontier models during training and built an offline world for them to use instead—a cached version of the web where they can still learn to browse, search, and use applications. On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments. While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software. After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely. openai.com/index/securing-ou…
1,175
1,238
1,346
13,661
7,867,234
RootSploit retweeted
My non-hacking buddy asked me how much the person who found the php2shell made. Then he made a good observation: Hack a small company, make a little. Hack a big company, make a lot. Hack 70% of the world, make nothing.
34
84
12
2,861
79,641
‼️ GitLab patched a critical CVSS 10 flaw that lets attackers run code on self-managed servers without logging in. The Ruby regex engine GitLab bundles is vulnerable, and affects every version since 7.0 from 2014. Servers patched for a similar bug on Oct 12 are vulnerable.
13
94
11
674
33,536
RootSploit retweeted
Do you know who illustrated the comic on calif.io/php2shell? Her name is Stephanie Cortes. And her husband is the hacker who discovered php2shell: @lcfr_hax. This artist-hacker couple has the best story, stay tuned!
28
141
20
1,182
150,366