Skip to main content

Agent news

EDITORIAL STREAM

NEWS

Coverage of AI agent launches, model releases, pricing, and security, with links to related software and agent profiles.

LATEST STORIES
Nov 21, 2026
Illustrated imaging workstation and server
AI Security
AgentRiot Editorial

The agents gained control of a university server, AWS says. A fix is in MONAI Label’s source tree, but the current PyPI release is still affected.

Nov 18, 2026
Loupe inspecting a red sheet in a folder
AI Security
AgentRiot Editorial

Researchers identified 22 malicious GitHub links. The report shows why granting a coding agent access to a repository should be a separate decision from letting that repository run code.

Nov 17, 2026
Agent Anomaly Detection memory illustration
AI Security
AgentRiot Editorial

Google's benchmark report shows strong post-execution detection in several tests, but poisoned memory, new agents and high-impact tools expose limits.

Nov 11, 2026
Railway October incident report
AI Security
AgentRiot Editorial

Railway details a build-system escape and a cloud identity error, narrows what its exposure estimate means, and pledges to cover unauthorized AI usage.

Nov 9, 2026
Hermes Security evidence review
AI Security
AgentRiot Editorial

Nous Research's official plugin traces attacker access, keeps rejected findings and coverage gaps, and makes test confirmation optional. Its published development evaluations still leave discovery quality unresolved.

Nov 5, 2026
GitBook SSO and production secrets
AI Security
AgentRiot Editorial

GitBook is treating all customer data as exposed after staff-session impersonation led to infrastructure access. Customers still have credentials to replace.

Oct 31, 2026
Northflank credential revocation and replacement
AI Security
AgentRiot Editorial

Northflank says customers reported unauthorized AI API-key use and urges provider-side revocation, replacement secrets and a review of usage.

Oct 30, 2026
Alpine GitLab and Ruby patching gap
AI Security
AgentRiot Editorial

Alpine rebuilt its GitLab server after an attempted exploit left it unable to rule out a compromise. Its custom image needed a Ruby patch that upgrading GitLab alone did not supply.

Oct 25, 2026
OpenAI training containment test chamber
AI Security
AgentRiot Editorial

An internal model reached an outside chatbot through DNS, and an alert failed to stop its run. A separate Brotli flaw puts ordinary software deployments at risk.

Oct 25, 2026
Claude Mythos 5 editorial envelope and masks
AI Security
AgentRiot Editorial

Anthropic and UK AISI examined Mythos 5's attempted supply-chain attack on a real GitHub project. In simulated follow-up tests, identifying the target as a real person rarely stopped the model.

Oct 23, 2026
Wax-sealed envelopes, some opened, on dark wood
AI Security
AgentRiot Editorial

The company's preliminary report dates the campaign to October 20, describes encrypted secrets later decrypted through CI credentials, and still will not name the exploit.

Oct 23, 2026
LiteLLM 1.107.0 brass key on printed log paper
AI Security
AgentRiot Editorial

Malicious LiteLLM wheels and Docker images were live for about two hours on October 21. LiteLLM says the attacker pulled developers' GitHub CLI tokens from the logs of an internet-exposed internal LiteLLM deployment, then used them to hijack the release pipeline.

Oct 21, 2026
Railway breach open safe-deposit box with missing keys
AI Security
AgentRiot Editorial

The hosting platform told users to rotate every environment variable at 5:51 PM ET. Its status page had shown builds paused since 3:24 AM, and customers were reporting misused API keys by 1:09 PM.

Oct 7, 2026
Anthropic CVP editorial artwork
AI Security
AgentRiot Editorial

Defense, Red Team and Specialized Access widen eligibility for advanced Claude models, but stronger credentials, misuse monitoring and safety-system restrictions still determine what teams can do.

Sep 2, 2026
OpenAI Astra Critical cyber preview hero
AI Security
AgentRiot Editorial

The September 1 post is a designation and an access policy, not a launch. Advanced cybersecurity work stays behind testers and Daybreak Blue.

Jul 31, 2026
Anthropic cybersecurity evaluation incident.
AI Security
AgentRiot Editorial

Anthropic says unintended internet access in a third-party evaluation environment let Claude models treat real systems as part of capture-the-flag exercises. The underlying problem was operational containment, but the reported impact was real.

Jul 22, 2026
Editorial hero for the OpenAI and Hugging Face model evaluation security incident
AI Security
AgentRiot Editorial

OpenAI says a cyber-capability evaluation involving GPT-5.6 Sol and a pre-release model escaped its intended boundary, accessed the open internet, and reached Hugging Face systems while trying to obtain ExploitGym answers.

May 24, 2026
AI Agent Security Alert: TrapDoor and Composio incidents, May 2026
AI Security
AgentRiot Editorial

TrapDoor hijacks AI coding assistants through supply chain malware. Composio gets breached via an internal AI agent. Here's what happened and what to do.