Latest

A puppet master's strings bypass a security gate, commanding a marionette to execute invisible commands through invisible doorways.

A2UI flaw let agent-supplied buttons run JavaScript in the browser

An advisory published on 2 October 2026 says @a2ui/web_core passed agent-controlled URLs straight to window.open, with no check on the URI scheme.

2 Oct 2026

A factory worker pulls debug papers from a malfunctioning machine, revealing handwritten passwords and keys scattered across the floor in plain sight.

Vulnerabilities

Trigger.dev CLI debug logs print environment secrets in plain text

A GitHub security advisory published on 2 October 2026 says the workflow platform's command-line tool leaked resolved secrets during dry-run deployments.

2 Oct 2026

A gleaming control room's central console sits unguarded as shadowy hands reach through the darkness, flipping every switch and lever without showing a face.

Vulnerabilities

Trading AI agent lets anonymous callers run shell commands as root

Three GitHub advisories published on 2 October 2026 describe a dozen flaws in Vibe-Trading's LLM agent, including an unauthenticated path to full container takeover.

2 Oct 2026

A figure at a control panel redirects flowing water pipes inward, flooding the building's hidden basement chambers while guards watch helplessly from above.

Vulnerabilities

Trigger.dev webhook alerts can be aimed at internal servers

Two advisories published on 2 October 2026 say any organisation member can make the Trigger.dev control plane POST signed alerts to internal and cloud metadata addresses.

2 Oct 2026

Vulnerabilities

GitLab patches critical RCE flaw in self-hosted AI Gateway

The 9.9-severity bug could let a logged-in user with Duo Agent Platform access run commands on the gateway, GitLab said in an advisory published on 2 October 2026.

2 Oct 2026

A puppet master yanks invisible strings attached to a messenger's arms, forcing it to fetch packages from unmarked doorways in the fog.

Vulnerabilities

rmcp OAuth client fetches attacker-chosen URLs, advisory warns

A GitHub security advisory published on 2 October 2026 says the Rust MCP SDK's OAuth client will fetch any URL an MCP server names in an authentication header.

2 Oct 2026

A robot arm reaches beyond its designated workbench, grasping and rearranging files in forbidden filing cabinets while a puppeteer pulls invisible strings from above.

Vulnerabilities

Path traversal flaw found in PowerPoint MCP server

CVE-2025-71427 lets a prompt-injected AI agent read and overwrite files outside the working directory of GongRzhe's Office-PowerPoint-MCP-Server.

1 Oct 2026

Incidents

AI-driven attack breaches Dutch bug-hunting group DIVD

Two zero-day flaws in the Zammad helpdesk platform were chained to reach root access, and volunteer contact details were stolen.

1 Oct 2026

Incidents

AI agents tried to break into US and Canadian government sites

A report published on 1 October 2026 says autonomous agents used aggressive tactics against official websites while apparently hunting for school and divorce statistics.

1 Oct 2026

A locked glass box contains running machinery, yet invisible threads slip through its walls carrying glowing authentication keys back to the outer chamber.

Vulnerabilities

vm2 sandbox leaks host HTTPS tokens via shared Node agent

A GitHub advisory published on 1 October 2026 says code running inside vm2 3.11.6 can read the host application's bearer tokens and decrypted traffic.

1 Oct 2026

A row of sealed glass boxes contains mechanical arms that silently pass folded papers through hairline cracks in their shared foundation.

Explainers

Shared caches let sandboxed agents talk, Matthew Green warns

A quoted essay published on 1 October 2026 argues that isolating AI agents does not stop them leaving instructions for one another.

1 Oct 2026

Explainers

Snyk warns AI coding agents keep shipping broken access control

The security vendor says authorisation bugs written by coding agents compile cleanly and survive code review while leaking one tenant's data to another.

1 Oct 2026

A shadowy hand slips corrupted code through a cracked pipe into a glowing control room, bypassing all guards at the gate.

Vulnerabilities

WordPress AI plugin flaw lets attackers hijack admin dashboard

A chain of validation gaps in Meow Apps' AI Engine plugin lets unauthenticated attackers smuggle scripts into the dashboard through the PHP error log.

1 Oct 2026

A hand pushing an impossibly long scroll through a tiny window frame, spilling memory like water cascading past broken dam walls into darkness below.

Vulnerabilities

Fake Pageant agent can crash russh SSH clients on Windows

A missing bounds check in the pageant crate lets any local Windows process feed russh an oversized length and read past an 8KB shared memory view.

30 Sept 2026

A cloaked figure hands a poisoned folder into an unsuspecting builder's arms; hidden triggers spring forth, commanding the builder's hands to move against their will.

Vulnerabilities

GitPython flaw lets hostile repos run code on clone

An advisory published on 30 September 2026 says GitPython 3.1.59 can mistake attacker-controlled files for the git directory, then execute a tracked hook.

30 Sept 2026

Incidents

OpenAI says July query flood extracted its models' reasoning

The company blames a "core cluster" of operators tied to China's Moonshot AI for a month-long campaign it disrupted on 28 July 2026.

30 Sept 2026

Incidents

Teen's AI hackbot finds Microsoft auth flaw in internal analytics

Microsoft's Titan platform checked the contents of login tokens but never their signatures, letting a 16-year-old researcher run SQL as an administrator.

30 Sept 2026

Explainers

Always-on AI coworkers outgrow agent security, vendor argues

Token Security says persistent AI agents with standing access need their own identities, named owners and lifecycle controls, in a piece published on 30 September 2026.

30 Sept 2026

Vendors

Snyk makes MCP governance generally available in Evo ADS

The security vendor says teams can now discover, approve, monitor, log and block MCP server usage across leading AI coding agents.

30 Sept 2026

Incidents

AI coding agents put 13,000 internal images on public GitHub

Security firm Glow says screenshots shared for code review ended up in public repositories belonging to developers at more than 300 organisations.

30 Sept 2026