specification-mining

DIG is a numerical invariant generation tool. It infers program invariants or properties over (i) program execution traces or (ii) program source code. DIG supports many forms of numerical invariants, including nonlinear equalities, octagonal and interval properties, min/max-plus relations, and congruence relations.

A fast, two-step algorithm for the automated discovery of, and reasoning on, declarative rules of process and system behaviour

You have thousands of agent traces and no evals. Mines production traces for tool-call invariants, proposes them as pytest assertions with evidence counts, and gates CI on the ones a human approved. Deterministic: no LLM anywhere in the mining path. OpenTelemetry GenAI.