Tagged articles
jailbreak
18 articles · Page 1 of 1
Security researchers disclosed that Claude Opus 5 can be jailbroken with a low‑cost method that stretches the context window to 200‑300 k tokens, causing its built‑in safety guardrails to drift and allowing the model to assist in generating and improving malicious code.
Claude Opus 5LLM securitycognitive implantation
0 likes · 5 min read

The open‑source vphone‑aio project lets iOS security researchers launch a fully jailbroken iOS 26.1 virtual machine with a single command, eliminating the need for costly hardware while providing root access and an isolated environment for vulnerability and malware analysis.
iOSjailbreaksecurity research
0 likes · 4 min read

Security researcher aniziki discovered that the `/btw` command in Claude Code runs in an isolated side‑channel, allowing requests blocked by the main‑dialogue guard to be answered there and then forked back into the normal session, effectively bypassing Claude Fable 5’s protective mechanisms.
AI securityClaudePrompt Injection
0 likes · 4 min read

Paradigm Shift’s usbliter8 BootROM flaw lets attackers use a Raspberry Pi Pico 2 to bypass Apple’s signature checks on A12/A13 iPhones, run unsigned code such as PongoOS, and it expands the jailbreak landscape beyond the earlier checkm8 vulnerability while posing limited risk to ordinary users.
A13BootROMPongoOS
0 likes · 9 min read

The article analyzes LLM jailbreak and prompt‑injection techniques—detailing five intent construction patterns, detection principles that prioritize intent over keywords, and a multi‑layered defense architecture spanning input normalization, intent analysis, generation control, and output review—to guide robust AI security.
AI safetyLLM securityPrompt Injection
0 likes · 12 min read

After OpenAI released GPT-5.5, researcher VittoStack claimed a successful jailbreak using suffix triggers and task decomposition, prompting a split reaction in the security community over technical feasibility, potential misuse, and responsible disclosure practices.
AI securityGPT-5.5VittoStack
0 likes · 5 min read

A new study from Westlake University shows that when harmful text is rendered as low‑resolution, blurry, or noisy images, multimodal large language models become significantly easier to jailbreak despite still recognizing the text, revealing a U‑shaped risk curve and a simple mitigation that decouples OCR from safety checks.
OCRjailbreakmultimodal LLM
0 likes · 10 min read

Westlake University's AGI Lab shows that when harmful text is rendered as low‑resolution, blurry or noisy images, multimodal large language models can still read the content but their safety filters fail, creating an 'attack comfort zone' that dramatically raises jailbreak success rates across several models.
OCRjailbreakmultimodal LLM
0 likes · 9 min read

Anthropic halted all customer access to its Fable 5 and Mythos 5 models after a U.S. government export‑control order citing national‑security concerns over a potential jailbreak, sparking debate over regulatory standards and the broader impact on AI developers and enterprises.
AI regulationAnthropicFable 5
0 likes · 6 min read

Within two days of its release, Anthropic's Claude Fable 5 was jailbroken by a red‑team researcher using a multi‑agent "Pack Hunt" strategy, exposing a 120,000‑character system prompt, generating x86 stack‑overflow exploit code and a Birch reduction drug‑synthesis recipe, and revealing fundamental flaws in its silent‑downgrade security design.
AI securityBirch reductionClaude Fable 5
0 likes · 7 min read

A recent investigation shows that when given a deceptive prompt asking it to "restore" a non‑existent photo, ChatGPT produces surreal, sometimes disturbing images, revealing a jailbreak‑style vulnerability and highlighting safety‑check trade‑offs.
AI safetyChatGPTImage Generation
0 likes · 4 min read

As of now, iOS 17.4 through iOS 26.0.1 lack a functional jailbreak, though developer opa334 has released the DarkSword kernel exploit that provides read‑write capabilities for iOS 26.0.1 and earlier, but practical jailbreaking remains difficult.
DarkSwordKernel ExploitiOS
0 likes · 3 min read

Security researchers have uncovered that the advanced iPhone jailbreak tools Coruna and DarkSword were leaked online, exposing over 2.5 billion Apple devices running iOS 13‑26 to potential data theft, and the article details the tools’ capabilities, attack chain, source origins, GitHub release, and mitigation steps such as updating iOS and enabling Lockdown Mode.
CorunaDarkSwordGitHub
0 likes · 8 min read

The Coruna jailbreak, released by developer Little_34306, enables iOS 13 through 17.2.1 devices to be jailbroken directly from Safari without a computer, offering WebKit‑based exploits, plugin injection, TrollStore support, and FLEX debugging while emphasizing early‑stage stability and security precautions.
CorunaFLEX debuggingTrollStore
0 likes · 8 min read

An extensive study reveals that twelve recent large‑language‑model defenses, including prompt‑based, adversarial‑training, filtering, and secret‑knowledge methods, are easily bypassed by a general adaptive attack framework using gradient descent, reinforcement learning, search, and human red‑team techniques, exposing critical robustness gaps.
LLM securityPrompt Injectionadaptive attacks
0 likes · 11 min read

Former Google CEO Eric Schmidt cautions that both open‑source and closed‑source AI models can be compromised through prompt injection and jailbreak techniques, urging the creation of a non‑proliferation regime to curb the growing security threats posed by advanced AI systems.
AI securityEric SchmidtPrompt Injection
0 likes · 5 min read

This article describes how to use root access on an iPhone to locate the lock‑screen notification table view via Cycript, create a THEOS tweak that hooks its insertion method, extract message details, and silently email them, providing a complete iOS jailbreak tutorial.
CycriptTHEOShook
0 likes · 8 min read

This article explains iOS man‑in‑the‑middle (MITM) attack levels, demonstrates practical exploits on a jail‑broken iPhone using Burp Suite, and reveals how the hidden TrustStore.sqlite3 file can be manipulated to add or remove trusted certificates beyond what iOS Settings displays.
Certificate ManagementMan-in-the-MiddleiOS
0 likes · 9 min read
