AMBIGUITY TESTING FOR AI-BUILT SOFTWARE
Your code can be green
while your spec is broken.
SpecFork gives the same ticket to independent AI agents. When valid implementations disagree, it shows exactly what your requirement forgot to decide.
Paste a product ticket⌘ ↵ to run
01 Find reasonable interpretations02 Build competing contracts03 Ask one decisive question
SpecFork does not ask whether a ticket sounds clear. It asks whether independent, reasonable implementations produce the same observable behavior.
Built with Codex + GPT-5.6