{"id":"CAP-1","type":"capability","title":"Evaluator recommendation observation","created":"2026-07-13","status":"offered with declared uncertainty","authors":["Upstream Zero"],"edges":[],"maturity":"experimental","body":"Structured observation of how AI evaluators surface, compare, validate,\nand eliminate a specific organization in matched buying scenarios: what is\nsaid, what is recommended, what evidence is requested, and where\nconfidence visibly breaks down.\n\n**Why this is marked experimental:** a capability cannot be marked\noperational until it derives from published, stable method (the build\nenforces this). M-1 is under active development and no observation method\nhas yet been published, so this capability is offered as what it is —\ncareful, honest, early-instrument work, with sampling conditions and\nvariance disclosed rather than hidden.\n\n**What this cannot tell you:** why the evaluator behaves as it does\n(mechanism), or whether its behavior will survive the next model update\n(see Q-3).","url":"/capabilities/CAP-1","machineUrl":"/objects/CAP-1","referencedBy":[{"from":"ENG-1","rel":"exercises"},{"from":"ENG-3","rel":"exercises"},{"from":"ENG-4","rel":"exercises"},{"from":"ENG-6","rel":"exercises"}],"_meta":{"site":"Upstream Zero — Commercial Evaluation Observatory","version":"0.1","note":"Claims are presented at their evidence tier; Narrated is the lowest. Verify by walking edges, not by trusting us."}}