Tagged articles
security research
43 articles · Page 1 of 1
The awesome-llm-apps GitHub repository offers 100+ end-to-end tested, CI-gated AI applications across 12 categories—from starter agents to multi-agent systems—compatible with major LLMs and licensed Apache-2.0, providing a graded learning path and a testbed for AI agent security research.
AI agentsApache 2.0CI/CD testing
0 likes · 9 min read

The article provides a comprehensive technical analysis of the critical cPanel authentication bypass vulnerability CVE‑2026‑41940, detailing its CRLF‑injection root cause, public PoC, large‑scale GitHub Actions abuse, ties to the “Sorry” ransomware, impact on hosted services, and recommended patching and mitigation steps.
CRLF InjectionCVE-2026-41940GitHub Actions
0 likes · 13 min read

The AretiQ security team disclosed a path‑traversal flaw in SharePoint Server's Upload.aspx page that lets an authenticated user with Contribute rights upload files to any library, and when PageParserPaths permits server‑side scripts, an ASPX webshell can be executed for full remote code execution, with CVSS 3.1 score 6.5 (Microsoft 8.2).
CVE-2026-45454SharePointUpload.aspx
0 likes · 14 min read

At Black Hat USA 2026, Zenity Labs disclosed the "PleaseFix" vulnerability that lets attackers hijack AI browsers via a malicious email without any user interaction, compromising Claude in Chrome, Gemini in Chrome, Perplexity Comet, ChatGPT Atlas, and Copilot Edge, and demonstrated further local‑host exploits and mitigation steps.
AI browsersBlack Hat 2026Intent Collision
0 likes · 8 min read

The Black Hat USA 2026 conference, the largest ever with more than 200 talks spanning AI security, zero‑day exploits, cyber‑warfare, and more, has uploaded all keynotes, main‑stage sessions and highlighted talks to its official YouTube channel for free viewing.
AI securityBlack Hatcyber warfare
0 likes · 6 min read

Over three months, the author collected thousands of Google API keys, built a scanner and an AI‑driven testing framework, uncovered multiple high‑value vulnerabilities across Google services, and earned more than $500,000 in bug‑bounty rewards, detailing each step and lesson learned.
AIGoogle APIsautomation
0 likes · 12 min read

HackerOne will mandate government‑issued ID verification for all bug bounty submissions from August 14, excluding minors and users on VPNs or rooted devices, sparking community outrage, raising privacy concerns, and potentially driving researchers toward gray‑market channels as regulatory pressure mounts.
HackerOneKYCbug bounty
0 likes · 9 min read

A security researcher discovered that the Claude Code CLI can be compromised by a crafted claude-cli:// URL that injects a malicious --settings flag, allowing arbitrary command execution without any user confirmation or warning dialogs.
Claude CodeCommand InjectionDeeplink
0 likes · 10 min read

The Certighost vulnerability (CVE‑2026‑54121) in Microsoft Active Directory Certificate Services allows a regular domain user to impersonate a domain controller, obtain a certificate, perform DCSync, and fully compromise an AD forest, with a publicly released PoC and detailed mitigation steps.
Active DirectoryCVE-2026-54121Certificate Services
0 likes · 12 min read

The usbliter8 BootROM exploit, disclosed by Paradigm Shift in June 2026, provides permanent code execution on A12/A13 devices, and the community has built a full jailbreak toolchain that can break into iOS 27 beta, though it requires RP2350 hardware and has notable limitations.
Apple A12Apple A13BootROM exploit
0 likes · 11 min read

A GitHub repository named cve-2026-poc-collection, maintained by researcher XZ1r0, aggregates twelve high‑severity CVE‑2026 proof‑of‑concept exploits across web, Linux, Windows and other categories, detailing their impact, implementation languages, and offering a search script to help security professionals quickly locate and analyze the code.
CVE-2026KubernetesPoC
0 likes · 9 min read

In July 2026, the AI model Kimi K3 from Moonshot AI discovered and generated working remote‑code‑execution exploits for four Redis versions within 27 minutes, detailing a double‑free Stream NACK flaw and a TDigest heap overflow, and providing open‑source PoC scripts and defense guidance.
AI vulnerability discoveryRedisexploit
0 likes · 9 min read

The article explains the Global Device Identifier (GDID) embedded in every Windows machine, how it is stored in the registry, used across Microsoft services, and how researchers can extract or attempt to spoof it with the GDID‑Extractor tool, revealing a four‑layer defense and privacy implications.
Device IdentifierGDIDTool
0 likes · 18 min read

The open‑source vphone‑aio project lets iOS security researchers launch a fully jailbroken iOS 26.1 virtual machine with a single command, eliminating the need for costly hardware while providing root access and an isolated environment for vulnerability and malware analysis.
iOSjailbreaksecurity research
0 likes · 4 min read

WordPress’s wp2shell vulnerability chain, disclosed by Adam Kues and Mustafa Can İpekçi, links CVE‑2026‑63030 REST route confusion with CVE‑2026‑60137 SQL injection to achieve pre‑authentication remote code execution via a single HTTP request, and the article details the exploit steps, detection scripts, and mitigation measures.
RCEREST APISQL injection
0 likes · 14 min read

The article contrasts how Microsoft’s Security Response Center praised researcher april_ivyyy with a CVE assignment while simultaneously deleting Nightmare Eclipse’s MSRC, GitHub, and GitLab accounts and issuing a legal threat, illustrating the divergent outcomes of vulnerability disclosure within the same ecosystem.
CVEGitHubMSRC
0 likes · 9 min read

Oversecured’s GitHub‑published report reveals 176 security flaws uncovered in Samsung’s factory‑installed Android applications between 2022 and 2025, spanning dozens of core apps, all of which have been patched after Samsung paid a total bounty of $163,475.
AndroidBountyOversecured
0 likes · 6 min read

A security researcher released over 42,000 words of OpenAI Codex desktop system prompts and tool definitions, revealing the AI's layered persona, dual‑channel workflow, skill‑calling mechanisms, tool set, and safety constraints, offering a rare window into AI agent design and security.
AI safetyCodexGitHub
0 likes · 8 min read

Released on July 8, 2026, xAI's Grok 4.5 outperforms competitors in software‑engineering benchmarks, consumes far fewer tokens, and has been praised by security researchers for its powerful vulnerability‑detection capability, while offering a $2/M‑input‑token pricing and 80 TPS inference speed.
AI ModelGrok 4.5benchmark
0 likes · 4 min read

The article reveals how a seemingly harmless "Claude Code Max Unlimited.exe" posted on The Pirate Bay serves as a honeypot, letting security researchers dissect malicious AI‑crack tools, reverse‑engineer them, and even turn the attackers' own samples against them.
AIClaudeMalware Analysis
0 likes · 8 min read

lpe-toolkit is a multi‑architecture Linux privilege‑escalation toolkit that bundles 24 exploits across seven CPU architectures, automatically detects kernel versions to filter patched bugs, offers pre‑compiled binaries for amd64, multiple execution modes, and practical use cases for red‑team testing, incident response, and security research.
CVELinuxlpe-toolkit
0 likes · 8 min read

Human researcher Jaeyoung Chung uncovered the Bad Epoll kernel race‑condition (CVE‑2026‑46242) that AI model Mythos missed, demonstrating that even the most advanced AI code‑auditing tools can overlook critical vulnerabilities while human intuition and deep analysis still dominate security research.
AI auditingCVE-2026-46242Epoll
0 likes · 9 min read

Security researcher tinopreter discovered a high‑severity web cache deception vulnerability on a self‑registration booking platform, where cached home pages leaked JWTs and, by exploiting SameSite Lax defaults with a top‑level navigation redirect, allowed account hijacking and earned a $2,000 bounty.
Cache BypassJWTSameSite Lax
0 likes · 12 min read

On June 9, 2026, security researcher Nightmare Eclipse released the RoguePlanet zero‑day exploit that leverages a race condition in Microsoft Defender to spawn a SYSTEM‑level command prompt on Windows 10/11 machines fully patched with the June updates, while also hinting at a possible BitLocker bypass.
BitLocker bypassMicrosoft DefenderRoguePlanet
0 likes · 10 min read

The article examines Google’s $500 reward for a high‑severity V8 out‑of‑bounds write vulnerability, tracing the historic decline of bug‑bounty payouts, the monopolistic role of major platforms, AI‑driven bug‑finding saturation, and the resulting challenges for security researchers both globally and in China.
AIGoogleV8
0 likes · 11 min read

An anonymous researcher with a legitimate MSRC account publicly released multiple Windows 0‑day exploits after his reports were ignored, leading to swift bans on GitHub and GitLab, sparking a heated debate over platform policies, coordinated disclosure failures, and the broader breakdown of the bug‑bounty ecosystem.
GitHubGitLabMSRC
0 likes · 9 min read

Researchers demonstrate that by exploiting NTFS junctions and symbolic links to create recursive directory structures—dubbed GhostTree—a normal user can generate billions of paths that cause EDR folder scans to enter infinite loops, effectively hiding malicious files from detection.
EDR evasionGhostTreeJunction
0 likes · 10 min read

When traditional decompilers output unintelligible C‑style code for Rust binaries, Oxidizer restores Rust‑specific structures, enums and macros, achieving up to 28% higher accuracy and 20% faster analysis, as demonstrated on 28 popular Rust projects and validated in user studies.
Malware AnalysisRustangr
0 likes · 7 min read

This article details a step‑by‑step penetration test on a university’s web platform, covering XSS file uploads, JWT tampering for arbitrary login, massive personal data leakage, SQL injection payloads, and the exposure of several AK/SK secrets, all with concrete screenshots and commands.
JWTSQL injectionXSS
0 likes · 5 min read

In April 2026 a trio of Windows Defender zero‑day bugs—BlueHammer, RedSun and UnDefend—were publicly disclosed after Microsoft’s Security Response Center repeatedly ignored the researcher’s reports, sparking a debate over responsible disclosure, corporate trust, and the urgent need to respect security professionals.
BlueHammerMicrosoftWindows Defender
0 likes · 6 min read

The RedSun proof‑of‑concept demonstrates that when Windows Defender detects a malicious file marked with a cloud‑based detection tag, it may rewrite the file to its original location instead of isolating it, allowing an attacker to replace system files and obtain administrator privileges.
RedSunWindows Defenderlocal privilege escalation
0 likes · 1 min read

The article introduces VeilShell, a Godzilla‑based reflection AES encryptor combined with a Data‑Flow Break and dynamic callback technique to generate PHP webshells, presents detailed evasion test results against Changting, Alibaba and VirusTotal scanners, and provides performance metrics and a GitHub link for acquisition.
AES encryptionData-Flow BreakDetection Evasion
0 likes · 3 min read

A 2021 CCC member survey and a large‑scale analysis of Exploit‑DB reveal that hackers predominantly use Shell scripts and Python, with notable overlap across both data sets, while language preferences shift over time toward Python and away from C, highlighting detection challenges and future trends.
Pythonexploit-dbhacking
0 likes · 11 min read


AntTechJun 8, 2025 · Information Security
The award‑winning ICSE2025 paper introduces CDFuzz, a lightweight custom dictionary technique that dramatically improves coverage and vulnerability discovery in grey‑box fuzzing, and a live session will dissect its design, validation, and impact on software security testing.
CDFuzzICSE2025grey-box fuzzing
0 likes · 4 min read

A security research team created a counterfeit VSCode extension in half an hour, demonstrated how easily malicious code can be injected and distributed through the VSCode Marketplace, and revealed that dozens of high‑value companies, security firms and even a national court were compromised, highlighting critical gaps in extension vetting and supply‑chain protection.
Malicious ExtensionSoftware Supply ChainVSCode
0 likes · 10 min read

Researchers from the Leiden Institute of Advanced Computer Science analyzed over 47,000 GitHub repositories, uncovering that many fake proof‑of‑concept exploits conceal malware, with nearly 5,000 repositories deemed malicious and detailed case studies revealing hidden trojans, Cobalt Strike tools, and stealthy information stealers.
GitHubSoftware Supply Chainmalware
0 likes · 6 min read


AntTechOct 11, 2022 · Information Security
The article provides a comprehensive technical walkthrough of the Hyper‑V DirectX component, detailing its architecture, virtual GPU configuration, attack surface, and step‑by‑step exploitation of four critical CVEs (CVE‑2022‑21918, CVE‑2021‑43219, CVE‑2022‑21912, CVE‑2022‑21898) with code snippets and debugging insights.
CVEDirectXHyper-V
0 likes · 11 min read


21CTOJun 17, 2022 · Information Security
MIT researchers have uncovered an unpatchable hardware flaw in Apple’s M1 chip’s Pointer Authentication Code, demonstrating a speculative‑execution‑based PacMan attack that can bypass the chip’s last line of defense and even compromise the kernel.
Apple M1MITSpeculative execution
0 likes · 5 min read

This article presents a collection of Spring Boot security vulnerabilities, detailing information leakage and remote code execution cases, with step‑by‑step exploitation guides and underlying principles, intended for security research and authorized testing only.
information leakageremote code executionsecurity research
0 likes · 2 min read

Researchers from 360 Netlab have uncovered RotaJakiro, a stealthy Linux backdoor malware first seen in 2018, which uses ZLIB compression, AES/XOR/ROTATE encryption, and hidden plugins to exfiltrate data and evade detection, with twelve functions yet its true purpose remains unknown.
C2Linux malwareRotaJakiro
0 likes · 4 min read


21CTOMar 20, 2021 · Information Security
Egyptian security researchers discovered a chain of flaws in TikTok’s Android app—including generic WebView XSS, Add Wiki Activity XSS, intent-based component launch, a Zip Slip in Tma Test Activity, and an RCE exploit—that can be combined to achieve remote code execution, and the report details TikTok’s remediation steps.
AndroidTikTokWebView
0 likes · 10 min read

This article expands on the author's BlueHat Shanghai 2019 presentation, summarizing Office‑related 0‑day and 1‑day vulnerabilities discovered between 2010 and 2018, categorizing them by component and type, and providing extensive references, analysis notes, and exploitation guidance for security researchers.
0dayCVEOffice
0 likes · 25 min read


ITPUBMar 22, 2016 · Information Security
The author details a seldom‑seen MSSQL blind injection discovered during a Google bounty, explaining why automated scanners failed, how manual testing with @@LANGID and @@TEXTSIZE revealed the flaw, and the proof‑of‑concept steps that ultimately earned a reward.
Blind InjectionGoogle BountyMSSQL
0 likes · 5 min read
