I'm not an engineer. I built a prompt-structuring tool anyway, using Claude Code — here's what actually went wrong

Short summary

A non-engineer shares an honest account of building Deep Prompt Studio, a prompt-structuring tool, using Claude Code. The first rule-based version was unusable; upgrading to the Claude API improved results but introduced a tedious verification loop since the author couldn't read the code. Key lessons: AI assistants often give false reassurance about code quality, and manual testing remains essential for non-technical builders.

• •Non-engineer built a prompt-enhancement tool using Claude Code
• •Rule-based first version failed; Claude API integration was necessary
• •AI assistants often give false reassurance — manual verification is critical

Generated with AI, which can make mistakes.