ci: master is red on duplicate-bodies -- seven widget-logic helper groups (digits_val, ws_*) are in no ledger #11013

Summary

Master of gHashTag/t27 is red on the duplicate-bodies check (workflow .github/workflows/dupe-ratchet.yml, tool tools/dupe_scan.py) and every open PR inherits it. The first brief for this fix named hex_digit (copy added by 6dd449e, PR #10278); that copy was already removed by #10296 (Closes #10294), so it is not the cause any more.

Measured on master 882fcff, run 39243687990: seven groups are "copied [2] time(s) and in no ledger": digits_val, ws_at, ws_back, ws_before, ws_fwd, ws_len_at, ws_len_before. They arrived with b92f6f1 (PR #10965, widget logic ports, epic #10937): lut_explain_logic and lut_treemap_logic share the ws_* byte scanners, fasm_skyline_logic and stat_card_logic share ws_len_at and ws_len_before, fasm_skyline_logic and lut_depth_logic share digits_val.

Each widget logic spec compiles to its own wasm module with module-level state and one parse per instance, so a cross-module use is a design change owned by the widget epic, not a one-line fix. The smallest correct change is to record the seven copies as deliberate with python3 tools/dupe_scan.py --bless in one commit, run on the Railway lab.

Boundary

• tools/duplicate_bodies_baseline.txt

User Scenarios

• Given master with seven unledgered duplicate groups, when the ledger records them and python3 tools/dupe_scan.py runs, then it exits 0.
• Given a PR that adds a new duplicate body, when the check runs after this change, then it still fails.

Requirements

• FR-001: The ledger MUST list exactly the seven groups above, each with count 2, and no other line may change.
• FR-002: The change MUST NOT add hand-written code in any language other than t27.

Success Criteria

• python3 tools/dupe_scan.py exits 0 on the branch, run on the lab.
• The duplicate-bodies check of the PR is green.
• 1 file changed, tools/duplicate_bodies_baseline.txt.