Varun Pratap Bhardwaj

Blog

From the Build Log

Thoughts on AI agent reliability, building open-source tools, and lessons from enterprise architecture at scale.

Technical deep-dives → Most technology-focused articles live on qualixar.com/research/blog — AI agent memory, MCP tooling, formal verification, and the Qualixar product engineering log. This blog is the personal research notebook.

·9 min read·philosophy

The Reaching: What AI Will Never Learn About You

My daughter reached for the light before she had a word for it. I build AI for a living — and that gesture undid everything I thought I knew about what a mind is.

consciousnessvedantaadvaitaawarenessfatherhoodai
·4 min read·philosophy

Guna Optimization: A Sāṅkhya Physics for Neural Networks

How Sāṅkhya's concept of the three cosmic Guṇas (Sattva, Rajas, Tamas) maps to hyperparameter optimization, learning rates, regularization, and convergence.

sankhyaoptimizationhyperparametersmathematics
·4 min read·philosophy

The Digital Tantra: Prakāśa and Vimarśa in the Agentic AI Stack

How the Kashmir Shaivism concepts of Prakāśa (luminous awareness) and Vimarśa (active self-reflection) map to the transition from static LLM weights to dynamic agentic loops.

shaivismagencyarchitecturemeta-cognition
·5 min read·philosophy

Pramāṇa for AI Agents: Indian Epistemology as the Original Reliability Engineering

How classical Indian validation systems (pramāṇa-vāda) map to modern software engineering for non-deterministic AI agent systems.

epistemologyverificationformal-methodsreliability
·7 min read·market-intel

Google Just Validated What We Built: Why Jitro Proves AI Agents Need Persistent Memory

Google's Project Jitro (Jules V2) is building a persistent agentic workspace with goals, insights, and history. This is exactly the problem SuperLocalMemory solved — locally, privately, and months earlier.

ai-reliability-engineeringsuperlocalmemorygoogle-julesai-agent-memorypersistent-ai-memoryai-agents
·6 min read·security

Operation Pale Fire: What Block's Red Team Proved About AI Agent Security

Block's security team ran a red team exercise against their own AI agent Goose and achieved full compromise. The findings reveal architectural vulnerabilities that affect every AI agent connecting to external tools via MCP.

securityai-agentsred-teammcpprompt-injectionai-reliability-engineering
·4 min read·enterprise

The EU AI Act Takes Full Effect August 2, 2026. Your AI Memory System May Not Be Ready.

The EU AI Act (Regulation 2024/1689) takes full effect August 2026. Most AI agent memory systems send data to cloud LLMs — creating compliance challenges that engineering alone cannot solve. Here is the architectural analysis.

eu-ai-actcomplianceprivacymemoryenterprise
·5 min read·research

SuperLocalMemory V3: Mathematical Foundations for Production-Grade Agent Memory

How information geometry replaced cloud LLMs for 74.8% accuracy on LoCoMo with data staying local. Fisher-Rao retrieval, sheaf cohomology, and Langevin dynamics explained.

memoryresearchinformation-geometryopen-sourceeu-ai-act
·4 min read·comparisons

SuperLocalMemory vs Mem0: When Zero-Cloud Beats Managed Memory (2026 Benchmark)

Benchmark comparison of SuperLocalMemory V3 and Mem0 on LoCoMo. SLM Mode A scores 74.8% without any cloud dependency — higher than Mem0's 58-66%. Factual analysis with trade-offs.

comparisonmem0benchmarklocal-firstlocomo
·7 min read·security

CVE-2026-25253: What the First Agent Software CVE Means for Developers

The first CVE for AI agent software was assigned in 2026. Understanding CVE-2026-25253, its implications for agent skill supply chains, and what developers should do now.

securitycvevulnerabilitysupply-chain
·10 min read·education

How AI Agents Learn from Their Actions: Behavioral Learning Explained

SuperLocalMemory v2.8 introduces behavioral learning — your AI tracks what works and what doesn't, surfacing better memories over time. No LLM inference required.

behavioral-learningv2.8ai-agentsmachine-learning
·9 min read·education

Why Your AI Agent Needs Memory Lifecycle Management

AI agents generate thousands of memories but never clean up. Learn how automatic lifecycle management keeps your AI fast and relevant without manual intervention.

lifecyclev2.8architecturememory-management
·8 min read·comparisons

Best Mem0 Alternative: Why Local-First AI Memory Wins

Comparing SuperLocalMemory vs Mem0 for AI agent memory. Why local-first beats cloud-dependent memory for developers who care about privacy and speed.

comparisonmem0local-firstprivacy
·7 min read·security

The ClawHavoc Attack: 1,200 Malicious AI Skills and How to Protect Yourself

In January 2026, the ClawHavoc campaign planted 1,200 malicious skills in AI agent marketplaces. Here's what happened, why it matters, and how to verify your agent skills are safe.

securityclawhavocagent-securityskillfortify
·8 min read·security

How to Secure Your MCP Server Configurations Against Supply Chain Attacks

Model Context Protocol servers connect AI tools to your system. Here's how to verify your MCP configurations are safe and lock them down against supply chain attacks.

mcpsecuritysupply-chainguide
·8 min read·tutorials

How to Give Cursor AI Persistent Memory in 5 Minutes

Step-by-step guide to adding persistent, local memory to Cursor IDE. Your AI remembers your codebase, coding patterns, and preferences across sessions.

cursormcpsetupguide
·10 min read·tutorials

Set Up an MCP Memory Server for Any AI Tool in 5 Minutes

Universal memory server using Model Context Protocol. Works with Claude Code, Cursor, Windsurf, VS Code, and 17+ AI tools. One memory across all your tools.

mcpsetupuniversalguide
·10 min read·education

Building a Universal Memory Layer for AI Agents: Architecture Patterns for Scalable State Management

Learn how to architect a memory layer for AI agents, covering episodic vs. semantic stores, hybrid retrieval, and multi-agent interoperability.

ai-agentsagent-memorystate-managementvector-searchhybrid-search
·14 min read·tutorials

How to Give Claude Code Persistent Memory (Free, Local, 5 Minutes)

Step-by-step guide to adding persistent memory to Claude Code using SuperLocalMemory V2. Free, local-first, works offline. Install in one command.

claude-codememorymcptutorial
·7 min read·updates

SuperLocalMemory v2.7: Your AI Learns You — Adaptive Learning That Runs Locally

SuperLocalMemory v2.7 introduces adaptive learning with local ML ranking — all 100% local. Your behavioral data never leaves your machine.

releaselearningprivacymlgdpr

Stay Updated

Weekly insights on AI agent reliability, new research, and tools I'm building. No spam, unsubscribe anytime.