Daily Edition · Page 5

THE DAILY INDEX

Technology, AI, software, security, crypto and business

ATTACKS & VULNERABILITIES

TLDR InfoSec

COYOTE TROJAN FIRST TO USE MICROSOFT UI AUTOMATION IN BANK ATTACKS

The Coyote banking Trojan is notable for being the first malware to utilize Microsoft's UI Automation (UIA) framework, transitioning from a proof-of-concept to actively targeting Brazilian banking and cryptocurrency users. It exploits UIA to steal credentials from browsers and can manipulate UI elements, such as address bars, to redirect victims to phishing sites, thereby complicating detection efforts. Akamai recommends monitoring for the presence of UIAutomationCore.dll in unknown processes and scanning for interactions with UIA named pipes as early indicators of infection.

THE CYBER DECEPTION MATURITY MODEL: WHERE DOES YOUR ORGANIZATION STAND?

This post divides cyber deception into 5 levels: pre-deception, where an organization does not employ cyber deception; experimental, where an organization deploys simple deception that only succeeds at catching automated scanners; operational, where an organization can catch human threats; strategic, where an organization's deception resources mimic authentic resources and mimic high-value targets; and optimized, where deception goes beyond detection and shapes adversary behavior. Each level includes key metrics to track as well as steps to advance to the next level.

CLOUD LOGGING FOR SECURITY AND BEYOND

This is a guide on cloud logging best practices for AWS, Azure, and Google Cloud. It emphasizes the importance of balancing security, compliance, and cost when selecting logging strategies. The guide distinguishes between control plane logging (default for management) and data plane logging (which requires manual activation for service activities), categorizing logs into audit, compute, network, secrets, storage, database, and Kubernetes types. Organizations should prioritize logging based on business, regulatory, and security needs, avoiding the costly "log everything" approach.

COMMON MISTAKES IN USING CVSS

Don't rely solely on CVSS scores for prioritizing vulnerabilities, as only 17% of critical ones are exploited, while attackers target lower scores. Organizations should implement Risk-Based Vulnerability Management (RBVM) that includes threat intelligence, asset importance, business context, and metrics like EPSS. Research shows EPSS prioritization targets 3% of vulnerabilities with 65% effectiveness, unlike CVSS, which requires 57% coverage for similar results, with only 4% effectiveness.

FLOWABLE'S SUMMER 2025 UPDATE INTRODUCES GROUNDBREAKING AGENTIC AI CAPABILITIES

Flowable released version 2025.1 on July 22 with AI capabilities, including four agent types: Utility, Document, Knowledge, and Orchestrator. It now supports integrations with AWS Bedrock, Azure Foundry, and Salesforce Agentforce, and features dynamic autonomy for agents switching between AI help and deterministic workflows based on task complexity. The enterprise 'AI button' enables real-time summaries, automated customer communication, and autonomous task completion.

CONTRACT LAPSE LEAVES CRITICAL INFRASTRUCTURE CYBERSECURITY SENSOR DATA UNANALYZED AT NATIONAL LAB

Critical infrastructure cybersecurity sensor data remains unanalyzed after the DHS contract with Lawrence Livermore National Laboratory ended this weekend, leaving the CyberSentry program unfunded. The program monitors threats in operational technology networks at power plants and factories, but the lab can't process the data without government funds. This occurs amid broader Trump administration budget cuts to CISA and expired threat hunting contracts.

TLDR IndexPage 5