TLDR InfoSec

TLDR INFORMATION SECURITY 2025-01-07

5 sections15 entrieslinks blockcomplete parse

Parse notes

  • Subscription/account footer truncated before parsefooter_truncated

01ATTACKS & VULNERABILITIES

WINDOWS 10 USERS URGED TO UPGRADE TO AVOID SECURITY FIASCO

ESET is urging Windows 10 users to upgrade to Windows 11 before Oct 2025, which is when Windows 10 reaches EoL and stops receiving security updates. Almost 63% of Windows users are still running Windows 10. Adoption has been slow due to missing features and the TPM requirement. Microsoft will be offering Extended Security Update plans for businesses and individuals.

bleepingcomputer.com2 min read

02STRATEGIES & TACTICS

HOW ATTACKERS ABUSE SERVERLESS SERVICES TO HARVEST SENSITIVE DATA FROM ENVIRONMENT VARIABLES

This post provides an overview of how attackers exploit environment variables from serverless devices post-exploitation. When secrets are stored directly in environment variables, they can be leaked by an attacker that can achieve code execution on the function. As permissions are generally provided to the serverless execution environment via environment variables, they can also be acquired by the attacker to pivot to other services in the account that the function can access. To mitigate these issues, utilize Secrets Manager services, audit functions, scope roles to least privilege, and alert on abuse.

permiso.io10 min read

HOW TO BUILD AN OFFENSIVE AI SECURITY AGENT

This post walks through the process of building a POC for an offensive AI security agent that was able to analyze a JavaScript file to discover API endpoints and detect a vulnerability. The author provided the AI agent with 4 “tools” to find endpoints, format an API request, execute the test, and analyze the results. The post also provides some lessons learned, such as the importance of effective prompting, the power of agentic workflows, and the strengths of Claude 3.5 Sonnet for code writing.

anshumanbhartiya.com7 min read

OPENAI'S RED TEAMING INNOVATIONS DEFINE NEW ESSENTIALS FOR SECURITY LEADERS IN THE AI ERA

OpenAI has released two papers showcasing its advanced approach to AI security through red teaming. The company employs external teams to find vulnerabilities and use automated reinforcement learning for attack simulation. Its human-in-the-middle approach combines human expertise with AI techniques, setting new standards for AI model safety.

links.tldrnewsletter.com7 min read

03LAUNCHES & TOOLS

FRIZBEE

Frizbee is a tool that takes a tag and returns a checksum.

github.comGitHub repository

ACCESS

Access is a tool that utilizes the Okta API to provide a centralized portal for employees to discover, request, and manage their access to internal systems.

github.comGitHub repository

04MISCELLANEOUS

FROM REMEDIATION TO MITIGATION: ADDRESSING INSECURE-BY-DESIGN FLAWS

Insecure-by-Design flaws are vulnerabilities inherent to the design of a system. They require a full redesign of the system remediate. These vulnerabilities should first be mitigated by implementing guardrails and monitoring tools. Once mitigations are in place, a slow sunsetting process can be started to reduce downstream impacts.

vectra.ai6 min read

NESSUS SCANNER AGENTS WENT OFFLINE DUE TO A FAULTY PLUGIN UPDATE

Tenable disabled two Nessus scanner agent versions due to a faulty plugin update that caused agents to go offline. The company released Nessus Agent v10.8.2 to address the issue and has advised customers to upgrade to this version or downgrade to 10.7.3 to resolve offline agent problems. Plugin feed updates were temporarily paused to prevent further issues.

securityaffairs.com2 min read