-
Lithoglyph must verify actual certificates before reporting them valid. Unimplemented builtin verifiers now refuse success; real verifier integration remains.
-
GNPL now has a runnable direct-evidence narration fragment: accounts, focalization, warrants, a limited rival relation and hypothetical withdrawal. Connect its imported-snapshot contract to Lithoglyph next, then extend warrant derivations and temporal semantics with their own proof obligations.
-
KRL’s draft and implemented fragment now have separate surface contracts and an executable conformance corpus. Resolve their syntax deliberately before claiming a unified four-operation implementation.
-
QuandleDB must distinguish candidates from checked equivalence. Stronger assurance requests now fail when only bucket membership is available.
-
VeriSimDB / VCL-UT: establish a mandatory, tested admissibility boundary for every operation advertised as checked, including mutations and transitions.
-
Lithoglyph / GNPL / Glyphbase: connect a real account operation and warrant failure through durable storage and the application surface.
-
QuandleDB / KRL: implement and check an explicit resolution witness; keep unresolved results and incomplete searches visible.
-
Vocarium / Hermeneia: agree and test one shared storage/warrant contract, including invalidation or withdrawal. Existing invoke/checker slices use distinct store implementations and do not prove this integration.
Maintain an acceptance example and a meaningful rejection/uncertainty example for every implemented boundary. Run them in the owning repositories' CI. Report local test results separately from remote CI, formal proof and deployment.
The public vocabulary is VCL-UT, GNPL, KRL and Hermeneia. Selection, filtering and reporting remain supporting capabilities; no extra public language is assumed.