tool · pip install awprism

awprism

Turn a failure into ranked hypotheses — and say what would confirm each one.

The problem

Debugging with an agent collapses onto the first plausible story, because nothing forces a second. The cost is not the wrong guess, it is the hours spent proving it — measured repeatedly here, where five hypotheses were spent on a service that was genuinely correct, and where a symptom named the wrong component so consistently that "the symptom names the INNOCENT service" had to be written down as a standing rule.

instead of trusting

the first explanation that fits

you check

the ranked alternatives, and the observation that separates them

Adopt it

pip install awprism

Hand it a failure and get back ranked candidate causes, each with the one observation that would rule it in or out.

Pairs with

awreason awrepl awgraph awdit

Pairing is composition, never dependency — awprism installs and runs on its own.

The rest of the Aither World

Every brick below is live, drawn from that repository's own published manifest.