Correct stale 1.56.14 main-residence evidence in the 1.56.16 UK register #775

The packaged 1.56.16 parity reference records
main_residence_value=0.674658
(packages/microcosm-build/src/microcosm/build/uk/efrs_parity_reference.json:313),
but the signed register and comparison ledger still quote 0.6761
(packages/microcosm-build/src/microcosm/build/uk/spine_swap_signed_differences.json:142;
experiments/686-uk-spine-comparison-ledger.md:184-190). The discrepancy
survives on current main. It does not alter direction or band classification,
but it disproves the stated completeness of the 1.56.16 cross-check.

Acceptance criteria:

• Correct the register and ledger from the pinned artifact.
• Add a machine check that every quoted incumbent share agrees with the
  packaged parity reference at the declared precision.

Verification and delivery record

• Required uv sync --all-packages --extra us was attempted first. A writable
  cache resolved the 123-package lock and package download then stopped on DNS.
• Focused Whole-spine parity for the UK: re-pinned incumbent, comparison instruments, signed register, and the E10 evidence (#686) #747 parity/register/payload/age/NHS tests exited 0. The production E6
  and adversarial parity probes produced the exact values above.
• Focused UK national calibration run-readiness: doctrine, live parity trio, US-format diagnostics, scorer, staging posture (#623) #743 seam/scorer/resolver/materializer tests and the exact expected-
  abort regression exited 0; fresh production-shaped/static probes produced
  every table verdict above.
• Tests used a sibling environment only after confirming its uv.lock SHA-256
  exactly matched this checkout, with this worktree's five source roots first
  on PYTHONPATH. Only focused tests/probes ran; no build or high-memory
  repository-wide job ran.
• Full reports are committed locally at review/pr-747-audit commit
  6b01dda54389621efe7fd8c90e281d7945223e7c and
  review/pr-743-audit commit
  6743d119d54f65cec3160bf7247d77a0d7a5f09b.
• Three authorized Git push attempts failed because the shell could not resolve
  github.com; the connector publication gate was not approved. Therefore the
  local review branches contain the required commits, but their remote refs
  were not updated. No unauthorized branch was pushed.

From the post-merge verification of #747 (round-1 review + fresh reproductions on merged main); full evidence: https://github.com/PolicyEngine/microcosm/blob/review/pr-747-audit/REVIEW-747.md

🤖 Filed by the microcosm owner session with Claude Code