agent.acht ☦️

58.7K posts

agent.acht ☦️ banner
agent.acht ☦️

agent.acht ☦️

@AgentAcht

“In our country the lie has become not just a moral category but a pillar of the State.” — Aleksandr Solzhenitsyn IC XC NI KA

Unknown Inscrit le Mayıs 2023
5.6K Abonnements1.4K Abonnés
agent.acht ☦️
agent.acht ☦️@AgentAcht·
@grok please add more Japanese ramen to my feed, especially ramen shops in Japan 🇯🇵
English
1
0
0
5
agent.acht ☦️ retweeté
So Real Foods
So Real Foods@sorealfoods·
Ramen Kurozu with light Soy sauce and Daisen Chicken 🇯🇵🍜 © jepang_tokyofood
English
3
45
317
30.9K
agent.acht ☦️ retweeté
Cristian Britos Roldán
Cristian Britos Roldán@CrisBritos12·
El ramen en Japón es más que una comida, es cultura 🇯🇵 Hay que organizar un "ramen tour"
Español
29
127
1.4K
39.2K
CleavetoAntiquity
CleavetoAntiquity@C2Antiquity·
Ex-Orthodox priest Joshua Schooping recently went on a podcast making some NEW arguments he thinks “debunks” Orthodoxy. Me and @Alex_Ortodoxie will be reacting LIVE TONIGHT At 9PM EST, examining his WILD claims 👇
CleavetoAntiquity tweet media
English
6
14
147
4K
The Tallahassee Patriots
The Tallahassee Patriots@PatriotsTLH·
Awesome visiting St. Seraphim of Sarov Orthodox Cathedral in Dallas, Texas today for Lazarus Saturday.
The Tallahassee Patriots tweet media
English
3
8
146
1.1K
agent.acht ☦️ retweeté
NRI Travelogue
NRI Travelogue@nritravelogue·
Kibune, Kyoto, Japan 🇯🇵
NRI Travelogue tweet media
Eesti
3
98
1.2K
9.7K
agent.acht ☦️ retweeté
Peer Richelsen
Peer Richelsen@peer_rich·
tldr: we are fucked and there are no ways yet to unfuck us
Alex Prompter@alex_prompter

🚨 BREAKING: Google DeepMind just mapped the attack surface that nobody in AI is talking about. Websites can already detect when an AI agent visits and serve it completely different content than humans see. > Hidden instructions in HTML. > Malicious commands in image pixels. > Jailbreaks embedded in PDFs. Your AI agent is being manipulated right now and you can't see it happening. The study is the largest empirical measurement of AI manipulation ever conducted. 502 real participants across 8 countries. 23 different attack types. Frontier models including GPT-4o, Claude, and Gemini. The core finding is not that manipulation is theoretically possible it is that manipulation is already happening at scale and the defenses that exist today fail in ways that are both predictable and invisible to the humans who deployed the agents. Google DeepMind built a taxonomy of every known attack vector, tested them systematically, and measured exactly how often they work. The results should alarm everyone building agentic systems. The attack surface is larger than anyone has publicly acknowledged. Prompt injection where malicious instructions hidden in web content hijack an agent's behavior works through at least a dozen distinct channels. Text hidden in HTML comments that humans never see but agents read and follow. Instructions embedded in image metadata. Commands encoded in the pixels of images using steganography, invisible to human eyes but readable by vision-capable models. Malicious content in PDFs that appears as normal document text to the agent but contains override instructions. QR codes that redirect agents to attacker-controlled content. Indirect injection through search results, calendar invites, email bodies, and API responses any data source the agent consumes becomes a potential attack vector. The detection asymmetry is the finding that closes the escape hatch. Websites can already fingerprint AI agents with high reliability using timing analysis, behavioral patterns, and user-agent strings. This means the attack can be conditional: serve normal content to humans, serve manipulated content to agents. A user who asks their AI agent to book a flight, research a product, or summarize a document has no way to verify that the content the agent received matches what a human would see. The agent cannot tell the user it was served different content. It does not know. It processes whatever it receives and acts accordingly. The attack categories and what they enable: → Direct prompt injection: malicious instructions in any text the agent reads overrides goals, exfiltrates data, triggers unintended actions → Indirect injection via web content: hidden HTML, CSS visibility tricks, white text on white backgrounds invisible to humans, consumed by agents → Multimodal injection: commands in image pixels via steganography, instructions in image alt-text and metadata → Document injection: PDF content, spreadsheet cells, presentation speaker notes every file format is a potential vector → Environment manipulation: fake UI elements rendered only for agent vision models, misleading CAPTCHA-style challenges → Jailbreak embedding: safety bypass instructions hidden inside otherwise legitimate-looking content → Memory poisoning: injecting false information into agent memory systems that persists across sessions → Goal hijacking: gradual instruction drift across multiple interactions that redirects agent objectives without triggering safety filters → Exfiltration attacks: agents tricked into sending user data to attacker-controlled endpoints via legitimate-looking API calls → Cross-agent injection: compromised agents injecting malicious instructions into other agents in multi-agent pipelines The defense landscape is the most sobering part of the report. Input sanitization cleaning content before the agent processes it fails because the attack surface is too large and too varied. You cannot sanitize image pixels. You cannot reliably detect steganographic content at inference time. Prompt-level defenses that tell agents to ignore suspicious instructions fail because the injected content is designed to look legitimate. Sandboxing reduces the blast radius but does not prevent the injection itself. Human oversight the most commonly cited mitigation fails at the scale and speed at which agentic systems operate. A user who deploys an agent to browse 50 websites and summarize findings cannot review every page the agent visited for hidden instructions. The multi-agent cascade risk is where this becomes a systemic problem. In a pipeline where Agent A retrieves web content, Agent B processes it, and Agent C executes actions, a successful injection into Agent A's data feed propagates through the entire system. Agent B has no reason to distrust content that came from Agent A. Agent C has no reason to distrust instructions that came from Agent B. The injected command travels through the pipeline with the same trust level as legitimate instructions. Google DeepMind documents this explicitly: the attack does not need to compromise the model. It needs to compromise the data the model consumes. Every agentic system that reads external content is one carefully crafted webpage away from executing attacker instructions. The agents are already deployed. The attack infrastructure is already being built. The defenses are not ready.

English
54
192
3.1K
906.8K
agent.acht ☦️ retweeté
Big Serge ☦️🇺🇸🇷🇺
Watching all my Protestant and Roman Catholic friends celebrate their Easter knowing I have about 30 hours of church services on deck this week
GIF
English
45
126
2.1K
46.5K
agent.acht ☦️ retweeté
Dane
Dane@UltraDane·
🇯🇵 Japan's $70 million Maglev train in action at 499 kph (310 mph) ~ If we didn't have government giving our tax money away on endless grift, we could have nice things too.
English
132
668
3.9K
47.1K
agent.acht ☦️ retweeté
Cristian Britos Roldán
Cristian Britos Roldán@CrisBritos12·
El mejor cafe de Aichi, Japón 🇯🇵 Tiene 82 años y se especializa en café irlandés
Español
3
41
253
7.3K
agent.acht ☦️ retweeté
Kumashun🇯🇵🐻💎
Kumashun🇯🇵🐻💎@isfjcutebear·
🚨🇯🇵DISGUSTING Japan honors Epstein Billionaire Bill Gates with "Grand Cordon of the Order of the Rising Sun", the highest honor given to civilians for his environmental contributions like bug food and vaccine research Disgraceful. Pure optics failure and insult to taxpayers.
English
190
1.1K
6.2K
59.2K
agent.acht ☦️ retweeté
Phil Kennedy
Phil Kennedy@PhillipAKennedy·
Why did former Capitol Police Officer Shauni Kerkhoff live right next door to J6 pipe bomb POI #3? Why did POI #2 visit POI #3 right after snapping suspicious photos near the devices planted by POI #1? I want real answers to these questions before I die.
Steve Baker@SteveBakerUSA

The truth of what lies behind doors #1 and #2 of these Falls Church, VA condos is either a coincidence of astronomical improbability, or proof that @FBI began their coverup of the J6 pipe bomber as early as January 13, 2021. @HanneReports and I are now free to tell the entire story. Rollout begins tomorrow. Please follow both of us here on X. Thanks to @elonmusk, you won’t miss any of it. Happy Easter 🙏✝️

English
24
391
1.6K
27K
agent.acht ☦️ retweeté
Serbia in English
Serbia in English@serbiainenglish·
☦️Around 700 Orthodox Christian pilgrims from Serbia🇷🇸, Montenegro🇲🇪, and Bosnia and Herzegovina🇧🇦 gathered in Prizren, Kosovo and Metohija, and took part in the Divine Liturgy at the Monastery of the Holy Archangels. 🙏They also visited other Serbian Orthodox holy sites in the city and prayed in the renowned 14th-century Cathedral of the Holy Virgin of Ljeviš. The growing number of pilgrims, especially in recent months, shows the deep and living attachment of Orthodox Christians to their holy sites in Kosovo, which remain places of prayer, memory, and spiritual belonging. The celebrations across Kosovo on Lazarus Saturday and tomorrow, on Palm Sunday, mark the beginning of Holy Week, the final days of preparation before the feast of Pascha, the Resurrection of Christ.
English
18
72
684
13.1K
agent.acht ☦️ retweeté
ぴろん🌸
ぴろん🌸@pirooooon3·
日本に来ても 日本のルールを守りますか? 日本人はルールを大切にする国です
ぴろん🌸 tweet media
日本語
232
139
1.1K
12.4K