Nothing is blocked

Work can be committed, pushed and shared in any state, and the records keep unreviewed work distinguishable from accepted work.

TRACE-OS has no commit hook and no gate. A check can fail, and says so, but it stops nothing. The governing-actions report reads each commit, for each principle, as recorded, missing, not applicable or unknown, and missing is a reported state, not a refusal.

Why not refuse?

A refusal rarely produces what it was meant to protect.

Take a dataset entering the project, one of the boundary transitions. TRACE-OS asks for an intake record saying where it came from and what was checked first. If nothing was checked, the record says “nothing stated”. Refusing the entry would not produce a check, only an entry nobody recorded. A person can still decide to return or refuse material at entry; that decision is recorded like an admission, and the tool removes nothing.

Review, likewise, happens afterwards, in batches, when an actor with the authority to accept is ready to decide.

Checking, acceptance and boundary control

These are three different acts. Checking produces evidence about whether a candidate meets a criterion. Acceptance records an authorized decision to accept, return or refuse it. Boundary control determines what may enter the project or advance. A failing check is evidence, not a decision, so TRACE-OS carries out the A principle, accept independently, by keeping the review states distinguishable rather than by blocking anything.

What it keeps instead

It keeps the states apart. Not reviewed, reviewed by a model, accepted, returned and refused each read as what they are. Model review remains advisory unless an authorized actor records an acceptance. Each decision names the work it covers, so accepting one unit of work does not accept later changes. See Review states.

Sharing changes no review status

Work may be shared when sharing is authorized, whatever its review state. The checks report what they find, and publication proceeds whatever they show. Sharing is a separate event from acceptance: it changes no review status, so unreviewed work stays visibly unreviewed.

What a record does not prove

“Recorded” means a record exists. It does not establish that a check was adequate, a reviewer independent, or a transition authorized in fact. The records make gaps visible; they do not establish that every required action occurred. What they offer is narrower: you can see what was looked at, by whom, against what, and what was not.