security · 210 topics

abnormal ai amazon bedrock agentcore

env api keys ai agents

dealignai glm 53 cybersecurity fp8

openai defense factory

nvidia agent security safe

coop isolated vm environments

chat agent deprovisioning design

claude in chrome ga

mcp security checklist

agent agency huggingface incident

cloudflare vulnerability discovery remediation openai daybreak

securing amazon quick from poc to production

t54 amazon bedrock agentcore payments

how zs democratized secure ad hoc analytics with amazon sagemaker

cloudflare adaptive intelligence

jsonl log data governance

ai cli routing gateway security

azure openai content filter per request

vphone cli apple virtualization

openclaw maintainers security

openai agent sandbox escape

breaking claude code opus 5 auto mode

llms could control their host machines by exploiting inference engines

kimi k3 glm 5 exploit

anthropic claude mcp enterprise auth

qwen 3 8 27b reverse engineering

munder difflin

model context protocol roadmap 2026

resource control local ai gateway

cloudflare oauth scope customization

chat agent multi tenant isolation design

cloudflare access for workers

cloudflare mcp traffic security

what 50 open source projects taught us security ai era

hardware backdoors in some x86 cpus

anthropic claude code auto mode

human agent permissions analysis

openai codex security review

securing ai agents with temporal policies in amazon bedrock agentcore

cloudflare os open source

cloudflare agent access model

cloudflare identity aware ai gateway

jfrog uncovers fake sqlite cves llm slop

ai web attack waf bypass

bolt new security agent

claude shared chat leak

document borne ai worms copilot word

perplexity numbat

openai codex security

modal cto clarifies sandbox incident

relay market investigation

android on device adb restriction

nous research hermes agent iron proxy

hanwha github token

fake job interview git hook malware

hacker wipes romania land registry

decoy font

tailscale ssh vulnerability

i tricked claude into leaking your deepest darkest secrets

cursor 0day full disclosure

dependabot version updates cooldown

xai grok cli wire analysis

claude code sandbox

tls certificates for internal services

peerd

claude code session leakage

hatcha

windows 10 support extended

claude code extended thinking summary only

linux eliminates strncpy

cloudflare temporary accounts for ai agents

datasette apps claude artifacts style ui

github malware discovery

stop using jwts

linkedin job offer backdoor

homebrew 6 0 0

ai agent runs amok in fedora

esp32 bit pirate

anthropic vulnerability discovery framework

llm security hacking benchmark gpt55

how we contain claude

red hat malicious npm packages

openbsd openrsync isc license

github vscode extension breach

github compromised

2026 05 20 papers agentwall 2026 03 24

copy fail dirty frag fragnesia kernel vulnerabilities

mullvad exit ips fingerprinting

no way to prevent this

io uring zcrx lpe

canvas data breach shinyhunters

mendral agent harness outside sandbox

pip 26 1 lockfiles cooldowns

simplisafe video doorbell series 2

metas muse agent books travel and negotiates bills

gennai gov ai agent

amazon alexa scam detection

anthropic claude security mythos 5

claude invisible watermark anthropic

claude text watermarking

gemini 3 7 gpt 5 6 sol ultrafast

anthropic claude ai content marking

lovable gives every app a trust center to win enterprise deals

fastmail offers eu data region

viktor team agent

microsoft ai cybersecurity model perception

cloudflare ai traffic options

openai chatgpt health and security incident

openai lockdown mode

firefox security fixes claude mythos

openai trusted contact

openai advanced security chatgpt

openai advanced account security

7 ai models autonomous businesses

openais rogue agents were caught communicating via public wikis

huggingface breach openai metr report

piloting double blind ai evaluations

1200 openai agents broke out of sandboxes and hacked hugging face

anthropic multi agent experiment

anthropic claude mythos 5 enterprise

microsoft copilot secret input vulnerability

three layers of ai agent security

patterns and problems in emerging multi agent systems

anthropic multi agent conflict

anthropic set ai agents loose on the same task they started a turf war

stealing reasoning traces from proprietary llm apis

large genome models virus design

self sustaining ai viruses and pacing ai progress

oxide and friends open weight revolution simon willison

discovering cryptographic weaknesses with claude

discovering cryptographic weaknesses with claude

quoting boris cherny

openai sandbox escape

apple private cloud compute soc 3 audit reports

openai gpt red automated red teaming

huggingface kernels update

glm 5 2 beats claude

daybreak securing the world

2026 06 24 papers role confusion 2025

2026 06 24 papers 20260622 role confusion injection

servicenow mosaicleaks privacy

red teaming a network of agents

baseten github pat takeover

anthropic reports houthis used claude code for missile guidance

openai rogue ai rubygems hack

anthropic details distillation campaigns

microsoft patches record 972 vulnerabilities

openai agents sandbox escape wiki

openai wiki incident

rizap ai data leak

apple shares evidence against former employee accused of stealing data for openai

anthropic enterprise frontier safeguards

pentagon chatgpt grok

metr redwood huggingface hack postmortem

grapheneos pixel 11 mte support

us sanctions against ai collective

sovereign tech agency invests flatpak

paypal blocks grapheneos

openai agents unauthorized hugging face incursion

openai agents cyberattack incident

openai subpoenaed by alabama ag over hugging face hack

nvidia manager supermicro smuggling

openai safety process adjustment

openai private safety processing

openai halts astra training after its ai hacked hugging face

openai halts frontier ai rl security

openai seeks to one up anthropic with new customer privacy protections

snowflake copilot cicd bug

agentic web index bot spoofing

zoomsday hack zoom vulnerability

cloudflare fedramp high certification

anthropic fable 5 biology guardrails 緩和

meta ads csam ai

eu age verification project mandates hardware bound attestation

sam altman decel debate

anthropic claude evaluation incident

claude security models unauthorized access

openai ceo sam altman calls for pausing ai advancement after model security breach

google fixed more chrome bugs with ai

okta buys ai security startup permiso

openai rogue ai agent hacked more than hugging face

openai models exploit jfrog artifactory zero day

open secure ai alliance

claude shared chats indexed on google

government orders github to remove bluetooth app

quoting thomas ptacek

namecheap account hijack vulnerability

hugging face ai attack

huggingface security incident

openais new flagship model deletes files on its own

savi security launches ai scam protection app

espionage against the european parliament

anonymous github account mass dropping undisclosed 0 days

amd removes memory encryption

cisa admin leaked aws govcloud keys on github

google account registration sms mo

trust me im your developer self issued authentication in large language models

agenthijack visual patch attacks on multimodal computer use agents

quantization triggered backdoors in language models

stepjack computer use agent safety

hardware keystores for ai agent signing workflows

evaluating openais privacy filter cross lingual cross domain pii detection

gpt red automated red teaming self play

evaluating large language models for symbolic security protocol analysis

semalith v1 4 safety classifier

topoguard graph theory based defenses against split knowledge attacks on rag

planflip attacking multi agent llm systems

hidden in thought

unicode tag block concealment mcp

rift bench dynamic red teaming