AI & LLM
AI and LLM covers generative models, agents, prompt injection, data security and new artificial intelligence tools. The collection connects capabilities, limitations, operational risks and their impact on technical work.
OWASP Launches OASIS: AI and AppSec Join Forces Against Open-Source Vulnerabilities
OWASP unveiled OASIS on August 26, 2026, a community project that pairs AI-generated patches with human AppSec validation to accelerat…
MikroTrick Hits RouterOS: Active Attacks and First Coordinated LLM-Assisted Disclosure
CERT Polska uncovers the MikroTrick chain — six vulnerabilities in MikroTik RouterOS, two critical for unauthenticated remote control.…
GPT-6 Astra Hits 100% on ExploitBench, but OpenAI Blocks Every PoC Request
OpenAI released GPT-6 Astra on September 4, 2026, achieving a perfect score on ExploitBench, the benchmark measuring the ability to tu…
AI Orchestrates Ransomware Attack in 10 Hours: Reaction Time Is Dead
A human attacker used frontier-model AI agents to compress a full ransomware operation from the typical two weeks down to roughly 10 h…
OpenAI: Autonomous Agents Used an Abandoned German Wiki for Three Months
Roughly 18,000 posts left on DSEwiki between May and July 2026. The GET/POST bypass mechanism reveals a gap between security testing a…
ASCII Smuggling: From AI Attack to 2.37 Million Phishing Emails
A technique originally developed for prompt injection against language models has spawned a financial phishing campaign exceeding 2.37…
From AI to Inbox: The 'Invisible' Technique That Flooded Corporate Email
Microsoft detected a campaign sending over 2 million messages per day by embedding invisible Unicode characters inside financial keywo…
OpenAI Commits $1 Billion to OT Defenders: Private Model Outpaces Federal Funding
OpenAI has pledged $1 billion in subsidized credits for its Daybreak for Frontline Defenders program, targeting under-resourced critic…
AI-Driven Attacks in Latin America: Unit 42 Exposes Two Operational Clusters
Palo Alto Networks' Unit 42 has uncovered two ongoing multi-stage intrusion campaigns across Mexico, Ecuador, and Brazil that leverage…
Autonomous AI Delivers Full Ransomware Attack in 10 Hours — and an 80-Page Audit for the Victim
Palo Alto Networks' Unit 42 has documented the first ransomware attack entirely managed by AI agents, completed in under 10 hours. The…
OpenAI's Astra Crosses the Critical Threshold: AI That Finds Zero-Days Without Guidance
OpenAI has designated Astra as the first model to reach the Critical threshold of its Preparedness Framework. The system discovers zer…
AI AppSec: Scanners Agree on Just 5%, Triage Costs $128,000
Contrast Security's AppSec Overflow 2026 report reveals three AI scanners testing the same codebase agree on only 5% of findings, whil…