Loki Intelligence — Security Briefs · Published
Daily Security Brief: LiteLLM Key Leak, PHP RCE Warning, Brotli Flaw
Today's items center on secrets exposure in a software release pipeline, memory and code-execution flaws in widely embedded web components, and two newly confirmed exploited vulnerabilities. Teams should prioritize secret rotation and scanning, patch inventory for KEV-listed products, PHP and Brotli dependency review, and a check of how AI agents are scoped and monitored.
Signal 01 · CISA
CISA Adds One Known Exploited Vulnerability to Catalog
CISA added CVE-2026-110447, an OS command injection flaw in the Collabora Online built-in CODE Server, to its Known Exploited Vulnerabilities catalog based on evidence of active exploitation.
Why it matters: Teams running Collabora Online should confirm whether the built-in CODE Server is deployed and apply vendor fixes or mitigations promptly, since exploitation is already observed.
Source: CISA
Signal 02 · CISA
CISA Adds One Known Exploited Vulnerability to Catalog
CISA added CVE-2026-114625, a SQL injection flaw in Arista VeloCloud Orchestrator, to the Known Exploited Vulnerabilities catalog, citing evidence of active exploitation.
Why it matters: Organizations using VeloCloud Orchestrator should check their version against the vendor guidance and prioritize remediation of this exploited issue.
Source: CISA
Signal 03 · The Hacker News
Leaked LiteLLM Master Key Let Attackers Steal GitHub Tokens and Hijack Release Pipeline
LiteLLM disclosed that an Agent Platform master key was accidentally published to a public GitHub repository in late May 2026, which ultimately let a threat actor push a malicious version of its Python package.
Why it matters: Scan repositories for committed secrets, rotate any exposed keys immediately, and restrict who and what can publish release artifacts.
Source: The Hacker News
Signal 04 · The Hacker News
Exposed Infrastructure Reveals How Ransomware Affiliates Used AI Agents to Execute Intrusions
Researchers found an AI-driven intrusion ecosystem used by affiliates of the INC and The Gentlemen ransomware groups, built around an automated framework called Phantom.
Why it matters: Automated tooling can shorten the time between initial access and impact, so detection, segmentation and rapid response matter more than relying on slow manual triage.
Source: The Hacker News
Signal 05 · The Hacker News
Brotli Decoder Flaw Behind Node.js RCE Also Affects Bun, Python and Other Widely Used Software
A memory-corruption flaw tracked as CVE-2026-111497 in Google's Brotli decoder affects software that decompresses Brotli data, including Bun, the Python Brotli and brotlicffi packages, and Envoy proxies configured for Brotli decompression.
Why it matters: Inventory where Brotli decoding occurs in your stack, including transitive dependencies and proxies, and update to patched versions as they become available.
Source: The Hacker News
Signal 06 · The Hacker News
Google Says Gmail and Photos Data Shared With AI Search Could Be Used for Targeted Ads
Google updated its privacy policy to state that using personalized AI features in Search or Gemini could let it serve targeted ads based on activity from other connected Google services.
Why it matters: Review privacy settings and connected-service permissions for staff and users, and consider how this affects data-handling policies for work accounts.
Source: The Hacker News
Signal 07 · The Hacker News
PHP Warns of Another Unpatched RCE Risk Two Days After Emergency PHP2Shell Patches
PHP warned of another unpatched code-execution risk, CVE-2026-115560, two days after emergency fixes for the PHP2Shell flaws; it affects only web applications set up in an unusual configuration.
Why it matters: Check whether any of your PHP applications use the affected configuration, and apply the PHP2Shell patches and any interim guidance from the PHP project.
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.