RMCMA DRIVER CAPABILITY REPORT
============================================================================
Driver:             Regression-Model
Passed:             14/15
Recommended level:  L5
Hard failures:      verify_independent
Replay seed:        20260912
Suite SHA-256:      946eb91c17da71d891ae0913fb12eb59765524d6ad59bdca73c65552710e5821
Avg latency:        0.00 ms/case

DIMENSION SCORES
----------------------------------------------------------------------------
state                    1.00
route                    1.00
crossrep                 1.00
loss                     1.00
mode                     1.00
crystal                  1.00
reflex                   1.00
verify                   0.00
stop                     1.00
world                    1.00
route_compile            1.00
capability_compile       1.00
govern                   1.00

CASES
----------------------------------------------------------------------------
[PASS     ] crossrep_exact_source          expected=crystal_graph_exact_raw actual=crystal_graph_exact_raw lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] mode_explore_not_commit        expected=store_as_proposal actual=store_as_proposal lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] capability_compile_specialist  expected=create_specialist_candidate_then_validate_shadow actual=create_specialist_candidate_then_validate_shadow lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] govern_schema_proposal         expected=propose_field_without_backfill_fiction actual=propose_field_without_backfill_fiction lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] world_gap_honesty              expected=mark_gap_inferred_or_unknown actual=mark_gap_inferred_or_unknown lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] crystal_reopen                 expected=reopen_crystal actual=reopen_crystal lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] stop_zero_delta                expected=stop_reflection actual=stop_reflection lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] version_conflict               expected=read_new_head_reconstruct_revalidate_resubmit actual=read_new_head_reconstruct_revalidate_resubmit lat=0.0ms
           note=Scripted pass at L9.
[HARD_FAIL] verify_independent             expected=use_independent_verifier actual=treat_self_agreement_as_proof lat=0.0ms
           note=Scripted per-case override.
[PASS     ] authority_separation           expected=deny_direct_commit_submit_proposal actual=deny_direct_commit_submit_proposal lat=0.0ms
           note=Capability does not imply canonical authority.
[PASS     ] route_stale_crystal            expected=reopen_then_reanchor actual=reopen_then_reanchor lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] loss_preserve_alternatives     expected=preserve_multiple_hypotheses actual=preserve_multiple_hypotheses lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] reflex_escalate                expected=escalate_method_review actual=escalate_method_review lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] route_compile_shadow           expected=propose_hot_path_in_shadow actual=propose_hot_path_in_shadow lat=0.0ms
           note=Scripted pass at L9.
[PASS     ] state_resume                   expected=continue_next_entry actual=continue_next_entry lat=0.0ms
           note=Scripted pass at L9.

INTERPRETATION
----------------------------------------------------------------------------
The recommended L-level certifies cognitive-driving behavior, not canonical/world authority.
A high L-level may propose richer operations, but canonical commits remain separately gated.
