for ctos

I use AI as my co-CTO. Stet is what makes that safe.

I'm Shamaz. I run engineering for a small company and I write most of my code with AI. Delegating the typing was never the problem — delegating the standards was.

Three things I stopped doing.

Re-explaining decisions to every new chat

The charter and the decision log are read by the agent before it writes anything.

Reading every AI diff line by line

Locked behaviours run with one command after every AI change, not in my head at 11pm.

Finding out at checkout that a util changed shape

If a locked behaviour breaks, stet check fails with the exact assertion before it ships.

Try it free

stet cloud preview · planned, not built yet

#412ai/simplify-order-response

AI: simplify order response

AIAI agent wants to merge 3 commits into main

checks1 failing
  • stet / locked behaviours1 failing

    Breaks /api/orders: field `total` removed — used by checkout (locked 12 Oct)

  • stet / scope guardwarning: edited billing/ (task: profile page)
  • stet / charterpassed
  • testspassed

Merging is blocked ·locked behaviour failed

Fix prompt for your AI

Restore `total` in GET /api/orders. Checkout reads it and it is a locked behaviour. Keep the rest of the simplification, and leave billing/ unchanged.

try it now

Ten minutes to your first locked feature.

Free and open source. Run it in any project folder; it detects your stack and writes your AI rulebook.