Built a multi-agent spec generator. Claude and Codex draft specs in parallel, cross-review each other's work, then a final arbiter produces the consensus spec.
Two models, two perspectives, better result.
Charlon
@charlon
- 0 Followers
- 1 Following
- 6 Posts
@rello For simple CRUD, you're right, one model is enough.
Where it shines: I was building a game engine with collision detection. Claude and Codex each kept introducing their own recurring bugs. Never the same ones.
By making them cross-review, each one catches the other's blind spots. The final spec avoids bugs both models individually keep making.
For "add a settings page" you don't need this. For "implement 2D physics with edge cases", it saves hours of debugging.
My grandma could push to prod on my SaaS. Without breaking anything.
Not because she can code. Because the stack won't let her break anything.
No SQL. No runtime errors. Migrations auto-generated, won't deploy if incomplete. Snapshots on every deploy. Missing env var? Won't deploy.
The language is Elm. The full-stack framework is Lamdera.
#ElmLang #Lamdera #FunctionalProgramming #WebDev #TypeSafety
@tngranados Cool post! I started with caffeinate + hooks too, but hit the 1h timeout limit. Built a native macOS app in Rust on top of that: hooks + pmset disablesleep + a daemon that monitors CPU to detect if Claude is actually working. Works with lid closed, and a background daemon auto-detects interrupted sessions (no hook for Ctrl+C yet sadly). https://github.com/CharlonTank/claude-code-sleep-preventer
@thepoliticalcat Thanks! Glad it's useful. No more coming back to a frozen terminal.
Made an open source tool for #ClaudeCode users on #macOS.
Problem: close your laptop lid → macOS sleeps → Claude stops mid-work.
Solution: a #Rust daemon that detects active Claude sessions and prevents sleep. When Claude finishes, normal sleep resumes.
One-click DMG install. MIT licensed.