[upstream #16433] 16433: judge override completeness after composing all priority layers

[berry-13]

Raised in the local cross-family review of LibreChat-AI#16433 (head f780dcd, finding local:f780dcd:L-002).

mergeConfigOverrides (packages/data-schemas/src/app/resolution.ts) validates each principal override layer, in priority order, on top of the librechat.yaml base plus the lower layers, and drops the nodes that leave that accumulated config invalid, including a required field that is still missing.

What happens: a low-priority layer that is incomplete on its own is dropped even when a higher-priority layer supplies the missing fields. For example, a role override endpoints.azureOpenAI.assistants: true over a base with no Azure endpoint is dropped, although a user override with higher priority supplies endpoints.azureOpenAI.groups. The result is safe (nothing invalid reaches consumers; the value beneath survives), but the lower layer's valid intent is lost.

Expected: completeness (required and related fields) is judged once all layers are composed, and only the nodes that leave the final composition invalid are dropped, while per-layer checks keep rejecting values that are wrong on their own.

[berry-13]

Related case from the local review of LibreChat-AI#16433 at c343969 (local:c343969:L-001): a field patch validated against librechat.yaml and the principal's own stored override returns 200, then merge-time validation drops it when another principal's lower-priority layer supplies a conflicting related field (for example cloudfront.imageSigning: "none" under a later cloudfront.requireSignedAccess: true). Composing layers once, as proposed above, is also what would let such writes be judged against the stacks they land in.