AI Security •
How to Detect npm Encrypted Loader Malware: 3.1M Downloads
Detect npm encrypted loader malware before it runs: the trigger matrix, AES-256-GCM payload, and IOCs behind mathmain's 3.1M weekly downloads.
9 min read
Read more →
Author
Umesh Malik
Software Engineer · AI Builder · Writer · Expedia Group
Software engineer writing about AI, Claude Code, LLMs, OpenAI, Anthropic, and developer tooling. 145 articles on AI engineering, production systems, and the tools shaping modern development. 5+ years at Expedia Group, Tekion, and BYJU'S.
Claude CodeAI Coding AgentsLLM EngineeringRAGFine-tuningOpenAIAnthropicNext.jsSvelteKitDeveloper Tooling
Google Search · Preferred sources
Prefer this site on Google
If you already read this writing, add umesh-malik.com as a Preferred Source. Google can then highlight it with a preferred badge in Top Stories, AI Overviews, and AI Mode — for you, not as a site-wide ranking boost.
Recent Articles 145 total
All articles →
AI Security •
Fix the libheif AVIF RCE: A CVSS 9.8 Bug in Sharp and Next.js
A CVSS 9.8 heap-buffer-overflow in libheif AVIF RCE (CVE-2026-84383) reaches Sharp, libvips, ImageMagick, and Next.js. Here's who's exposed and how to patch.
8 min read
Read more →
LLM Engineering •
How to Fix Multi-Agent LLM Latency: Text Handoffs Cost 2.5x
Multi-agent LLM latency isn't random: an ICLR 2026 paper measured the cause — text handoffs cost 2.5x more time and 3-5% less accuracy than skipping text.
11 min read
Read more →
AI Coding Agents & DX •
How to Scope an AI Agent for Reverse Engineering: 4 Weeks, Not Years
How to scope an AI agent for reverse engineering: one failing test beats a full spec. A GPU driver project shipped in 4 weeks, not years.
10 min read
Read more →
AI Coding Agents & DX •
How to Decide When to Give an AI Agent Autonomy: 3:1 Test
How to decide when to give an AI agent autonomy: chip design keeps 3 to 5 spec engineers per builder before trusting it — the same ratio tests your task.
10 min read
Read more →
LLM Engineering •
How to Migrate a Large System Prompt to Ollama Without Breaking It
Migrate a large system prompt to Ollama and it can burn 14% of a 65K context window before the first turn. What breaks, why, and the fix that worked.
8 min read
Read more →