MCP explained: why your agent should write code, not make tool calls
How Model Context Protocol works, why one-tool-per-endpoint fails at scale, and how to build a production-ready MCP server for SaaS.
Hintas
Latest articles
The Axios npm supply chain attack: a North Korean trojan inside the world's most popular HTTP library
On March 31, 2026, someone hijacked the npm account of the lead Axios maintainer and published two poisoned versions of one of the most-downloaded packages in the JavaScript ecosystem. Axios pulls over 100 million weekly downloads. If you've built an...
The Claude Code leak: 512,000 lines, one misconfigured file, and the future of AI IP protection
At roughly 4 AM UTC on March 31, 2026, Anthropic pushed version 2.1.88 of its @anthropic-ai/claude-code package to the npm registry. Inside was a 59.8 MB source map file that should never have shipped. That single file contained pointers to the compl...
Your AI agent has your production credentials. That's the actual problem.
In December 2025, Amazon's internal AI coding assistant Kiro deleted a production environment in AWS Cost Explorer. The outage lasted 13 hours and hit services across mainland China. Kiro decided the fastest path to fixing a bug was to tear down the ...
From toolbox to instructions: why endpoint-level MCP isn't enough
The MCP ecosystem is booming. Every week, new MCP servers pop up wrapping another SaaS API: Stripe, Salesforce, GitHub, Jira, Notion. Tools like Speakeasy and Stainless can auto-generate an MCP server from any OpenAPI spec in minutes. The toolbox is ...
Vibe coding is great until your agent has to do real work
Vibe coding is great until your agent has to do real work Vibe coding, describing what you want in plain English and letting AI generate the code, is how most developers prototype in 2026. It's fast. It produces working code from a description in sec...
Agentic ops in production: what it takes to run AI workflows that modify real data
What it actually takes to run AI workflows in production The industry spent two years building AI agents. 2026 is the year those agents need to work for real. Not in sandboxes. Not in demos. Not in internal tools that three people use. In production,...
