Thoughts on software engineering, AI-assisted development, and building things for the web.
Working on multiple things with Claude Code can quickly become hectic, overcrowded and hard to follow.
Putting everything in order is a key on-going process. The most obvious solution is to keep a record of the specifications in the codebase itself. A single source of truth about how code components should be working individually and together.
The more granular you're, the easier is would be for the AI agent to load specs and effectively load context and requirements when running tasks for you.
Starting a new codebase with Claude Code or another AI coding agent, requires a lot of fine tuning to make long-term, production ready code development consistent.
The main entry points for day-to-day interactions with an agent within a project are CLAUDE.md and custom bot commands. In combination with specs, these become the bread and butter of your daily coding activities.
Using tools like Claude Code on a daily basis can be quite amazing, however, It happens that you cannot fully trust it to securely do you bidding, without taking any risk, even though Anthropic is supposedly doing their best to prevent it as much as possible.
Multiple users have experienced sudden loss of data for various reasons, but nonetheless due unexplica ble shell command hallucinations. I'm not waiting! This is how I secure my dev environment before it's too late!
How I researched, designed, coded, and deployed my blog in a single day—heavily assisted by AI tools.
My AI-powered coding workflow with Claude Code, Copilot, and ChatGPT.
