Desk · Labs & reverse engineering

Research

Technical deep dives, reproducible tests, and tool evaluations.

Articles: 177
Topics: 20
Page: 8 / 23

Data graphic: GLM-5.3-Flash's FP8 weights take 328 GB for 320B parameters 1.025 bytes per parameter . A bar chart of effective bytes per parameter shows FP8 and INT8 at 1.0 to 1.05 and the 4-bit formats lower: INT4 GPTQ/AWQ 0.52 to 0.60, NVFP4 0.56, MXFP4 and EXL3 0.53. Worked examples: 328 GB fits four 128 GB DGX Sparks 512 GB , Qwen3.8-Flash in NVFP4 at about 135 GB is 7 GB over one Spark, and a projected 4-bit GLM-5.3-Flash of about 170 GB is an estimate, not a file size.

NVFP4 vs FP8 vs MXFP4 vs EXL3: The Quantization Format Guide for Local LLMs

The same 320B model is 328 GB in FP8 and near 135 GB at 4 bits. Bytes-per-parameter math, per-format tradeoffs, and how to size a box that actually fits.

Sep 12 · 13 min

FIPS 140-2 sunset infographic detailing the September 21, 2026 transition to Historical status, procurement blocks, and FIPS 140-3 requirements.

stableResearch Fips 140 2 Sunset

FIPS 140-2 Sunset September 2026: The Migration Guide to FIPS 140-3

Prepare for the September 2026 FIPS 140-2 sunset. Learn the procurement impacts, technical breaking changes, and exact steps to migrate stacks to FIPS 140-3.

Sep 12 · 8 min

Architecture flow diagram of EU Cyber Resilience Act Article 14 showing PSIRT notifications routing through the ENISA Single Reporting Platform to national CSIRTs within 24-hour and 72-hour statutory deadlines.

stableResearch Cyber Resilience Act

EU Cyber Resilience Act Article 14: Navigating Mandatory Vulnerability Reporting Requirements for Enterprise PSIRTs

Explore the EU Cyber Resilience Act Article 14 mandatory 24-hour vulnerability reporting rules, ENISA platform architecture, and PSIRT compliance steps.

Sep 12 · 9 min

Data graphic: CVSS 9.8 Critical for CVE-2024-34102 in Magento / Adobe Commerce; attack path from XXE reading env.php to stolen crypt.key, forged admin tokens and rogue admins, beside four eviction steps rotate crypt.key, purge oauth tokens, audit admin_user and core_config_data, purge CDN caches .

stableResearch Cosmicsting

Beyond the Patch: Evicting CosmicSting (CVE-2024-34102) Backdoors, and Why StyleSmuggler (CVE-2026-75650) Is a Separate Zero-Day

How to detect and evict CosmicSting (CVE-2024-34102) backdoors from Magento and Adobe Commerce, plus a correction: StyleSmuggler is a separate zero-day, CVE-2026-75650.

Sep 12 · 9 min

Cybersecurity dossier diagram illustrating the GitLab unauthenticated arbitrary file read vulnerability, detailing path traversal exploitation, key exfiltration from gitlab-secrets.json, and the impact of forged administrative session tokens.

stableResearch Gitlab Vulnerability

GitLab Unauthenticated Arbitrary File Read Exposes Critical Secrets: Full Blast Radius and Incident Recovery Runbook

Critical GitLab arbitrary file read flaw leaks server secrets. Learn how to hunt logs, stop admin session forgery, and execute an incident recovery runbook.

Sep 12 · 9 min

Threat dossier infographic summarizing CVE-2026-20131 exploitation in Cisco Secure FMC by Interlock ransomware, showing CVSS 10.0 severity, affected releases 7.2 through 7.7, and pivot risks.

stableResearch Interlock Ransomware

Interlock Ransomware Weaponizes Critical Cisco FMC Flaw: Why Standard Patching Fails

Interlock ransomware targets Cisco FMC flaw CVE-2026-20131 to breach networks. Discover why standard patching fails and how to audit and remove backdoors.

Sep 12 · 7 min

stableResearch Deepseek V4 1 Flash

Run DeepSeek V4.1 Flash on Three DGX Sparks

A stability-first SGLang deployment serves DeepSeek V4.1 Flash across three NVIDIA DGX Sparks: 750k KV cache, ~1600 tok/s prefill, ~38 tok/s single-stream decode.

Sep 12 · 2 min

Technical architecture diagram of a hardened DeepSeek harness runtime showing the reasoning engine, context manager, rootless Docker sandbox boundary, and security controls mitigating CVE-2026-82533.

stableResearch Deepseek Harness

Hardening DeepSeek Harness: Architecture, Setup Tutorial, and Runtime Security

Deploying DeepSeek-R1 agent runtimes? Learn how to harden the harness architecture, configure local inference, mitigate CVEs, and enforce sandbox security.

Sep 12 · 7 min