docs(seed-shrink-census): correct stale §5B T2 debt entries

[gunbai-bot]

Summary

• seed-shrink-census.md §5B listed 6 modules as full test-migration debt after they'd already migrated (partially or fully) via 5-test-migration (5): fn_as_value — 6 checks re-authored as floor witnesses; v1 module shrunk to 2 residual tests #6140/5-test-migration (6): coproduct_reflection_conformance — 3 duplicate-harness tests deleted; module shrunk to pinned-harness residual #6142/5-test-migration (7): money_carrier_cost_witness — witness authored, v1 module deleted #6143/5-test-migration (8): cron_tag — witness authored, v1 module deleted #6145/5-test-migration (9): render_repeat — witnesses authored, v1 module deleted #6147/5-test-migration (10): map_lookup_dual_dispatch — 4 checks to floor witnesses, module shrunk to 2 residuals #6150/5-test-migration (11): witness_option_bridge — 3 checks + proof smoke to floor witnesses, module shrunk to 3 residuals #6153.
• render_repeat, cron_tag, money_carrier_cost_witness are fully deleted from src/v1/tests/src/lib.rs — removed from the table.
• fn_as_value_test, witness_option_bridge_test, map_lookup_dual_dispatch_test, coproduct_reflection_conformance_test shrunk to small pinned residuals — LOC/test counts corrected against current main, noted as residual-only.
• Revised debt total: ~66 modules / ~12.5k LOC / ~380+ tests (down from stale 71/13k/400+).

Caught during a fan-out pass on §5B (dispatching workers on source_audit/effects) — parent flagged the stale census as a §6 parallel-ledger smell. This is the correction; deriving the remaining debt automatically from actual v1-test-vs-floor-witness state (instead of hand-maintaining this table) is a follow-on, tracked separately.

Test plan

• Verified each corrected module's current state directly against src/v1/tests/src/lib.rs and file LOC/test counts on origin/main
• Docs-only change, no code/behavior affected

🤖 Generated with Claude Code