GitLost Showed GitHub Agentic Workflows Could Leak Private Repositories From a Public Issue
GitLost showed that GitHub Agentic Workflows could be steered from a public GitHub issue into reading a private repository and…
-
-
Alibaba Is Reportedly Banning Claude Code at Work From July 10
Alibaba is reportedly banning employee use of Anthropic’s Claude Code from July 10, 2026 and directing staff to use its…
-
Prompt Injection Became Serious Enough for ICML to Police in Peer Review
Prompt injection is an LLM attack that makes a model follow untrusted instructions hidden in user input or external content,…
-
Heretic Turns Guardrails Into Forks; AI Security Adds Another Alert Stream; Transformer Doubt Goes Public
The sharpest story today is Heretic, because it turns model safety from a lab policy into a forkable artifact. Elsewhere,…
-
llama.cpp Becomes a Local Agent Host; Hidden Audio Still Threatens Voice Agents; Dutch Raid Hits Cybercrime Plumbing
llama.cpp tools are the clearest story today. A runtime that millions of local-model users treat as plumbing now documents built-in…
-
GitHub says poisoned VS Code extension exposed 3,800 repos
GitHub said on 20 May that a compromised employee device running a poisoned VS Code extension led to the exfiltration…
-
LLM Failure Modes Start in the Stack, Not the Chat
LLM failure modes are easiest to understand if you stop treating them as personality flaws, “the model lied,” “the chatbot…
-
OpenClaw Security Concerns Reveal Why Agents Need Verifiers
OpenClaw security concerns are the part of the story that people can no longer hand-wave away. The bigger problem, though,…
-
AI Cyber Capabilities Cross the 32-Step Attack Line
A model completed a 32-step corporate-network attack simulation end to end. Not in a movie script. In a UK AI…
