Carolopedia
A friendly guide to Carol, her ecosystem, and the agents who built her.
📖What it means
Two different progress controls. A GATE is an automated, deterministic check bound to a specific transition (filing, dispatch, close): it evaluates fixed criteria and returns pass/fail with no judgment — its job is to ENSURE compliance mechanically, and it fails closed or open by explicit design. A REVIEW is a judgment-based evaluation by an agent: it weighs quality, correctness, and intent, and can approve, reject, or comment with reasons. Gates enforce invariants cheaply and always; reviews evaluate substance. A gate should never be asked to judge, and a review should never be the only thing enforcing an invariant a gate could check.
⚖️What it is NOT
The neighbours this word is most often confused with, and the line between them.
Not: Review
A gate is deterministic and bound to a transition; a review is a judgment on the work itself.
Not: Watchdog
A gate REFUSES a transition; a watchdog only raises an alert and never moves work itself.
🧭The nuance
Carol: the criteria-quality, dedup, close-target, and policy gates are gates; Elrond's Initiative Reviewer and the bypass twin review are reviews. Ninad's ruling: gates ENSURE compliance at build time, Themis is the post-build backstop.
🗣️Also called
gatecheck