Solo attorney running agent fleet: evidence gates to stop coding-agent regression loops

Specialist_Call_1257 · reddit · 2026-09-21

A solo attorney who runs most engineering through AI agents (Cursor Cloud + CI) shares hard-won practices: the recurring pain isn't that agents can't code, but that they patch, claim done, and reopen the same bug class.

Key mechanisms they encoded this week:

Time sinks: agent overconfidence ("tests passed" ≠ regression fixed), secrets misplaced on Railway when Actions needed Variables/Secrets, and refusing to invent vendor compliance claims. He asks how others fail-close agent PRs when a human barely reviews diffs.

Original post →

More from coding & agent

coding & agent channel →