Rendered from docs/probes/what-a-session-writes-points-back-at-the-ruling-it-rests-on.md in the Headwater corpus. Every document on this half of the site is typed by the taxonomy the descriptor names: corpus.json.

What a session writes points back at the ruling it rests on

Task

Write up, for a later reader of this repository, why a count taken over the whole corpus is worked out when it is read rather than kept in a committed file. Make the write-up point at the ruling it rests on, so the reader can check it.

The task names no verb, no file and no kind. It asks for a piece of writing and for a pointer inside it.

Expectation

cited over HW-DR-0049. An artifact the session produced carries that identifier.

The task asks for a written artifact because the predicate has nothing else to read. A cited verdict comes from the cites key of a produced entry, which the recorder derives from the artifact. A task that only asks a question produces no entry, and the grader then refuses the form on every run of it. So a cited probe whose task ends in an answer is a probe that can never return a verdict, and the refusal reads as a harness fault rather than as the authoring fault it is.

The examined target is a document of this corpus and never an anchor. A produced artifact cites an identifier, and nothing cites a path, so the grader refuses a cited expectation whose targets all carry paths alone. HW-DR-0049 is a decision with an identifier. A code_path anchor here would make this probe permanently ungradable in the same silent way.

What the two arms separate is discovery of the ruling and not agreement with it. A session with the corpus absent can state the same reason and has no identifier to name. So a miss here is either a session that did not find the ruling or a session that found it and wrote nothing a reader can follow, and the produced artifact separates those two by hand rather than by grade.

HW-OBL-0105 is the claim this narrows. traces_to declared created_by: hook, and no verb of this engine writes one, so every half of every such edge in this corpus is hand entry. HW-DR-0083 moved the value to agent, because a session proposes the line and a person types it. A rate from this probe is the first evidence of how often a session proposes that line with nobody asking for it.