Recording an Agent Run at Its Non-Deterministic Boundaries Turns an Incident Into a CI Regression Test

arXiv 2609.20625·medium signalXBlueskyLinkedInCopy link

Chronicle (arXiv 2609.20625, 17 Sep 2026) records an LLM agent run as immutable envelopes at each non-deterministic boundary, then replays it. Its cut-point replay serves a chosen subset of boundaries from the record while executing the rest live against new code, which converts a recorded failure into a test that runs in CI. On six recorded failures, recording added 23 microseconds per crossing (0.008% of an assumed 300ms model call), full replay issued zero model calls and was bit-stable across 20 repetitions, and cut-point tests failed on faulty code and passed on guarded and benign changes for all six.