Loki Intelligence — Security Briefs · Published
Daily Security Brief: WordPress KEV, LLMjacking, Barracuda
Active exploitation of a WordPress Core race condition and rising LLMjacking costs show that both widely deployed web software and cloud AI credentials are being targeted. Teams should prioritize patching the WordPress flaw, tightening AI API key and spend controls, and reviewing third-party scripts and edge devices.
Signal 01 · Cloudflare Security
See the agents, set the terms: new visibility and control for personal agents
Cloudflare describes how personal AI agents blur the line between automated and human web traffic. It outlines tools that give site owners visibility into these agents and control over their access.
Why it matters: Site owners should decide deliberately how agent traffic is identified, rate limited and permitted, rather than treating all traffic as either human or bot.
Source: Cloudflare Security
Signal 02 · CISA
CISA Adds One Known Exploited Vulnerability to Catalog
CISA added one vulnerability to its Known Exploited Vulnerabilities catalog based on evidence of active exploitation. The entry is CVE-2026-119977, a time-of-check time-of-use race condition in WordPress Core.
Why it matters: Inventory your WordPress installations and apply the vendor fix promptly, since KEV listing signals real-world exploitation.
Source: CISA
Signal 03 · The Hacker News
CISA Confirms Hackers Exploiting WordPress Core Flaw During Software Upgrades
The Hacker News reports that CISA confirmed exploitation of the WordPress Core flaw tracked as CVE-2026-119977, rated high severity with a CVSS score of 8.1. The coverage ties the issue to software upgrade activity.
Why it matters: Prioritize updating WordPress Core and check that managed hosting or auto-update mechanisms have actually applied the fix.
Source: The Hacker News
Signal 04 · BleepingComputer
Google: LLMjacking attacks rack up $1 million in cloud AI charges
BleepingComputer reports on Google findings that LLMjacking, the hijacking of large language model access, is producing large fraudulent cloud bills. Some organizations reportedly lose more than $100,000 per day.
Why it matters: Set budget caps, usage alerts and tight scoping on cloud AI credentials so abuse is detected and limited quickly.
Source: BleepingComputer
Signal 05 · The Hacker News
Barracuda Flaws Could Be Chained to Bypass Authorization and Gain Root Access
Barracuda disclosed two vulnerabilities in its products: a pre-authentication authorization bypass and a command-related flaw. Chained together, they could let an attacker gain root access to a susceptible device.
Why it matters: Apply Barracuda's fixes to affected appliances and limit exposure of management interfaces to untrusted networks.
Source: The Hacker News
Signal 06 · The Hacker News
Google Warns LLMjacking Attacks Are Costing Some Enterprises Up to $1 Million in Cloud Charges
Google warned that threat actors are hijacking enterprise access to LLMs to consume expensive compute and AI services fraudulently. It describes LLMjacking as resource hijacking adapted for LLMs, with costs reaching up to $1 million for some enterprises.
Why it matters: Treat LLM API keys and cloud AI permissions as high-value secrets, rotate them regularly and monitor for unusual consumption; authorized testing of AI agents and APIs can help validate those controls.
Source: The Hacker News
Signal 07 · The Hacker News
APT29-Linked Hackers Hijack Web Analytics Script to Deliver iOS and Android Exploits
A Russia-linked actor known as APT29 is reported to be behind a watering hole campaign aimed at iOS and Android users. It involves compromised websites, fake domains and a supply chain compromise of the Ukrainian analytics service Plerdy.
Why it matters: Review third-party scripts on your sites, apply integrity controls and content security policies, and keep mobile devices updated.
Source: The Hacker News
Brief sources
Related briefs
- What Is AI Agent Security Testing?
- AI Red Teaming vs Traditional Pentesting
- MCP Security Testing Checklist
Relevant Loki service: Orvyn — AI agent security private preview.