Harshal Dev
Cloud & AI Engineering Blog
Deep dives, production architectures, and automation guides focused on building resilient systems and cognitive AI agents.
-
Building Self-Hosted AI Agents with OpenClaw
A comprehensive guide to running personal, secure AI agents in your own infrastructure using the open-source OpenClaw platform.
-
Build an Airtable MCP Server: Turning Your Database into a Dynamic Agent User Interface
A complete developer's guide to building a Python-based FastMCP server that connects AI agents directly to Airtable, transforming low-code bases into real-time agent control panels.
-
Atlassian Jira Rovo MCP vs. Custom MCP Servers: A Comprehensive Integration Guide
Understand the architectural differences between native Atlassian Rovo remote MCP servers and custom FastMCP Jira implementations, complete with client integration code and best practices.
-
Gemini and Claude: How to Unleash and Combine the Ultimate LLM Superpowers
A comprehensive guide on the contrasting and complementary strengths of Google's Gemini and Anthropic's Claude. Learn how to combine their powers for maximum productivity in coding, automation, SRE, and complex reasoning tasks.
-
An LLM is Nothing Without Skills: Building Autonomous Systems with agent.md & MCP
Why raw LLM reasoning falls flat in real-world automation, and how structuring agents using modular skills (plugins) and stateful configurations (agent.md) unlocks actual production-grade autonomy.
-
Driving Deterministic AI Automation with Local Agent Configs and Custom Skills
Learn how to transform standard LLMs into highly disciplined, autonomous software engineers using version-controlled agent configurations and reusable skills.
-
Automating the Web: Building Playwright-Powered MCP Servers Controlled by Agent Skills
A complete developer's guide to wrapping browser automation scripts as Model Context Protocol (MCP) tools and governing them securely using local agent personas.
-
Designing Multi-Agent Handoff Systems Using Markdown Blueprints
A deep dive into setting up supervisor-worker collaboration patterns using structured agent.md configurations and handoff skill guides.
-
Designing Executable Verification Checklists in Agent Skills
How to eliminate deployment bugs by embedding self-verifying test commands and assertion steps directly within local agent skill definitions.
-
Version Controlling AI Agent Behaviors in Team Workflows
Why you should store agent instructions inside Git, tracking behavior changes over commit history, and standardizing prompt templates for team alignment.
-
Optimizing AI Agent Context Windows via Structured Discovery Skills
How to design context discovery instructions to help local code agents navigate repositories, find config parameters, and minimize token costs.
-
Building Systematic Error Recovery Blueprints in AI Agent Skills
A guide to defining systematic troubleshooting routines, exception handlers, and auto-recovery rules within local agent configuration profiles.
-
How to Build and Configure Custom Agent Skills for AI Coding Assistants
A comprehensive developer's guide to defining custom AI agent personas, task workflows, and repeatable skills using version-controlled markdown configurations.
-
The Leap to Autonomy: Transitioning from Gemini CLI to Antigravity 2.0
A deep dive into the transition from standard Gemini CLI to the agentic capabilities of Antigravity 2.0, exploring core features, autonomous workflows, and the future of AI pair programming.
-
How to Build a Custom YouTube Transcript Fetcher and Video Downloader in Python
A complete step-by-step engineering guide to programmatically fetching subtitles, handling auto-generated caption fallbacks, and downloading media using yt-dlp—no YouTube API keys required.