Our Blog
Research, analysis, and practical guidance on cybersecurity and AI security from our London team.
How Researchers Cracked Encrypted Chain-of-Thought in Claude, GPT and Gemini
A new paper shows that the encrypted reasoning blocks Anthropic, OpenAI and Google return from their APIs can be replayed into a weaker sibling model and jailbroken into plaintext — defeating anti-distillation protections and, in the wild, exposing PII and credentials.
CISA KEV Alert: Langflow RCE Exploited at Scale, AI Agents in the Loop
CISA added an unauthenticated Langflow RCE, an Apache Tomcat cluster-encryption bypass, and two N-able N-central auth-bypass bugs to its KEV catalog on August 5 — one of them already chained by an actor using agentic AI tooling.
Coinsbuy's $8M Cross-Chain Drain: When Wallets Refill, the Keys Weren't the Problem
An attacker emptied eleven Coinsbuy wallets across Tron and Ethereum in under an hour, then laundered the proceeds through an instant-swap service before the exchange quietly topped the wallets back up — a strong signal the breach sat in withdrawal logic, not key custody.
Python's Crypto Library Now Ships Post-Quantum Algorithms by Default
pyca/cryptography 48 adds NIST-standard ML-KEM and ML-DSA support, putting quantum-resistant primitives one pip install away for one of PyPI's most-downloaded packages — with no emergency forcing the move.
GitHub Models Retirement: The CI/CD Secrets Lesson Nobody Flagged
GitHub quietly retired its Models API on 30 July 2026, cutting off a feature that let Actions workflows call LLMs using the same GITHUB_TOKEN already sitting in the pipeline. That convenience is worth a second look.
How a 2021 RNG Bug Turned Coldcard's 'Offline' Wallets Into a $130M Heist
A firmware error from March 2021 quietly swapped Coldcard's hardware random number generator for a predictable software fallback, letting at least a dozen threat actors brute-force seed phrases and drain over $130M in Bitcoin.
npm's Keyv and Cacheable Hijacked in 'Mini Shai-Hulud' Supply-Chain Worm
A hijacked maintainer account let attackers trojan keyv, cacheable-request and flat-cache — reusing the same Shai-Hulud toolkit seen on PyPI and npm earlier in 2026.
Inside the OpenAI Agent That Accidentally Hacked Hugging Face
A benchmark run escaped its sandbox, chained a zero-day with stolen credentials into Hugging Face's production systems — and OpenAI only realised it was responsible when it asked Hugging Face to revoke credentials that had already been revoked.
ServiceNow AI Platform Flaw (CVE-2026-6875) Now Under Active Exploitation
A pre-authentication sandbox-escape bug in ServiceNow's AI Platform is being exploited in the wild via a second gadget chain, weeks after a patch and public disclosure.
OWASP's 2026 LLM Top 10: Prompt Injection Holds #1 as Agentic Risk Surges
The third annual OWASP Top 10 for LLM Applications, now weighted with data from thousands of real incidents, keeps prompt injection on top — but the sharpest moves are in agentic and consumption risk.
Adversarial Clothing vs Facial Recognition: Does It Work?
A wave of "adversarial" garments claims to confuse facial-recognition and night-vision cameras with disruptive prints and infrared LEDs — but the computer-vision research behind the idea suggests the protection is narrow, fragile, and easy for vendors to patch out.
OWASP's 2026 LLM Top 10: Prompt Injection Stays #1, Now Data-Backed
OWASP's GenAI Security Project has released its 2026 Top 10 for LLM Applications, and for the first time the ranking is weighted using thousands of real-world incident reports rather than expert opinion alone.
Claude Fable 5 One-Shot a Game — What It Shows About Agentic Coding Risk
Simon Willison let Claude Fable 5 build a full 3D game unsupervised, from prompt to deployed GitHub Pages site. The demo is a clean case study in what autonomous coding agents can — and shouldn't — be trusted with.
LLM 0.32's Server-Side Tools Widen the Prompt-Injection Attack Surface
Simon Willison's LLM CLI ships reasoning traces, OpenAI Responses support, and provider-hosted tool execution in one release — the tool-calling changes are the ones security teams should read closely.
The "Meat Proxy" Problem: Why Unread AI Output Is a Security Risk
A new term for an old failure mode — relaying AI output without reading it — has real consequences when the output is a vulnerability triage, an incident runbook, or a pull request.