Neural decompilation candidate selection closes gap on obfuscated binaries
Generate-then-select pipeline with execution-guided selection reaches 88.8% pass@1 on obfuscated inputs.
Stories per day · 30 days
2 RSS
Relatedadversarial-evaluation1appsec1corellium1decompilation1guardsquare1jailbreak1masvs1mobile-security1neural-networks1nowsecure1reverse-engineering1
19 stories, newest first · one card per story
Stories tagged #obfuscation
Obfuscation leads jailbreak success across seven evaluated LLMs
Obfuscation averaged a 54.3% attack success rate, and self-reminder prompting cut ASR to 10% and 5% on two models.
12 Best Mobile Application Security Testing Tools Compared (2026): Features & Pricing
A 2026 roundup compares 12 mobile AppSec tools for testing, shielding, and device virtualization.
4
An Evaluation of the Semantic Understanding Capabilities of Large Language Models for Web Attack Payloads
PayloadSemBench evaluates how deeply large language models understand 240 web attack payloads across four semantic layers.
GlassWorm Supply Chain Attack Hides Malware Inside VS Code Color Themes
GlassWorm-linked malicious VS Code color themes on Visual Studio Marketplace and Open VSX shipped staged JavaScript loaders that pulled C2 addresses from Solana memos.
8
SC WordPress Malware: A Self-Healing Mesh of Loaders, Drop-Ins, and a Blockchain-Controlled Backdoor
Sucuri documents SC, a self-healing WordPress malware mesh persisting across files, database, and shared memory, regenerating within seconds of cleanup.
Context-Aware Functional Modeling for Android Third-Party Library Detection
LibFan detects partial Android libraries under heavy obfuscation, beating prior detectors on a new benchmark.
Vidar Malware Rewrites Its Obfuscation With Every Build to Make Detection Harder
Vidar infostealer rewrites its string obfuscation on every build using custom virtual machines and stream ciphers, blunting static detection of the credential thief.
2
Vidar Uses Custom Bytecode Interpreter and ARX Stream Ciphers for Per-Build String Obfuscation
Vidar stealer now hides strings with a custom bytecode VM and per-build ARX stream ciphers, weakening static detection.
3
Vidar Adds Virtual Machine and Custom Stream Ciphers For String Obfuscation
ThreatLabz shows Vidar 2.x and 3.x hide strings with a bytecode virtual machine and custom stream cipher.
4
LausivLoader analysis, or how to pass data between malware stages, (Thu, Sep 17th)
SANS dissects a LausivLoader JavaScript malspam sample that passes staged payload file paths to PowerShell via process environment variables.
1
RiskChainBench: A Benchmark for Obfuscated Platform Message Restoration and Evidence-Grounded Web Investigation
RiskChainBench benchmark evaluates models restoring obfuscated abuse messages and investigating risky websites, revealing execution failures as the main bottleneck.
PuzzleMask: Abusing Plain Prose as a Covert AI Attack Vector
Check Point details PuzzleMask, a plain-prose technique that bypasses LLM gatekeeper policy checks, letting hidden payloads reach target models unreviewed.
1
Understanding the Security Boundary of Obfuscation-based On-Device LLM Protection
Researchers formalize obfuscation primitives for TEE-protected on-device LLMs and show a Collapse attack breaks ArrowCloak, TSQP, and LoRO, then extend the boundary.
JSCeal Hides Crypto Malware in V8 Bytecode
Check Point details JSCeal, a crypto-stealing malware delivered as compiled V8 bytecode, and built a decompiler pipeline to expose its capabilities.
3
Once popular for attacking AI, ASCII smuggling is embraced by spammers
Spammers adopt ASCII smuggling—invisible Unicode tag characters—to evade email filters, with Microsoft Defender detections spiking to 2.5 million per day.
ReCAST: Restoration-aware Cascaded Stage-wise Training for Obfuscated SMS Risk Classification
ReCAST distills a large teacher model's de-obfuscation ability into smaller models for robust classification of obfuscated Chinese SMS fraud messages.
JavaScript obfuscation: From party trick to phishing kit
Cisco Talos explains JavaScript obfuscation techniques used by phishing kits and the approaches researchers use to reverse them.
New Agent Tesla Malware Variant Boosts Evasion Capabilities
KnowBe4 documented an Agent Tesla v4 campaign using emoji-based code obfuscation to evade detection.
5