ToolRiotAll guides
Everything we've published.
Newest first. Each guide says which kind of coverage it is: how we test AI tools.
- Guide
We Gave Two AI Coding Agents the Same Spec. Here’s How Differently They Worked.
We gave Prime Agent and Claude Code the same coding spec, pinned the same model, and checked both results independently. Both passed. Their workflows looked very different.
- Tested and selected
We Built a Release Queue With GitHub Spec Kit and Claude Code
We gave GitHub Spec Kit and Claude Code a real publishing problem: a local release queue with dependency rules, holds, and duplicate protection. Every rule held.
- Guide
We Used Crawl4AI to Turn Hacker News Into JSON
We asked Crawl4AI for the headlines and links on Hacker News and got 30 records we could verify, plus a rendered page and screenshot from the same test.