refactor(main): extract shard s2 union (aux_config_cmd + model_picker) from main.py

[andrexibiza]

Godfile kill — hermes_cli/main.py shard s2

Extracts the shard s2 union from hermes_cli/main.py into two focused modules (5x2x3 blind-witness extraction, waves 1-3 verified):

• hermes_cli/aux_config_cmd.py (+396) — auxiliary config command cluster
• hermes_cli/model_picker.py (+522) — model picker cluster

Verbatim method bodies; module-level test constants stay; regression tests shipped.

Line math: 23 added / 859 deleted in hermes_cli/main.py (moved into the new modules). hermes_cli/main.py re-exports the moved names so all existing callers resolve identically.

Testing

• tests/hermes_cli/test_w1_s2_extraction.py (+86) shipped
• Seam identity verified: moved names resolve through main.py re-exports

Related #78791 #78792 #77376 #77746 #77748 #77751 #77752 #77756 #77759 #79066 #79067 #79068 #79069 #79070 #78689 #78690 #78691 #78692 #78693 #78694 #78695 #78696 #78697 #78698 #78699 #78700 #78701 #78702 #78703 #78704 #78705 #78706 #78707 #78708 #78709 #78710 #78711 #78712 #78713 #78714 #78715 #78716 #78717 #78718 #78719 #78720 #78721 #78722 #78723 #78724 #78725 #78726 #78727 #78728 #78729 #78730 #78731 #78732 #78733 #78734 #78735 #78736 #78737 #78738 #78739 #78740 #78741 #78742 #78743 #78744 #78745 #78746 #78747 #78748 #78749 #78750 #78751 #78752 #78753 #78754 #78755 #78756 #78757 #78758 #78759 #78760 #78761 #78762 #78763 #78764 #78765 #78766 #78767 #78768 #78769 #78770 #78771 #78772 #78773 #78774 #78775 #78776 #78777 #78778 #78779 #78780 #78781 #78782 #78783 #78784 #78785 #78786 #78787 #78788 #78789 #78790

[spfcraze]

This was generated by AI during triage.

Summary:
The title and commit message say this PR extracts txn/task-link mixins from kanban_db.py (shard s2), but the diff contains no change to hermes_cli/kanban_db.py — the actual content is the hermes_cli/main.py shard s2 extraction (aux_config_cmd + model_picker) plus two contributor-email files.

Problems:

• The PR diff adds hermes_cli/aux_config_cmd.py and hermes_cli/model_picker.py and edits hermes_cli/main.py and tests/hermes_cli/test_w1_s2_extraction.py; "kanban_db" appears nowhere in it, and hermes_cli/kanban_db.py (which exists at origin/main with its own test suite) is untouched.
• The commit 99e83f6d — whose message the title matches — carries exactly this main.py shard s2 content, not a kanban_db.py extraction.
• The body's "Line math: 1029 added / 859 deleted in run.py" names run.py, which is not among the files this PR changes.

Solution:
As a title matching the diff: "refactor(main): extract shard s2 union (aux_config_cmd + model_picker) from main.py" — which is what this PR actually contains.

---

Checked against 99e83f6 — the tip of gfg/main-extract-s2-w1a when this was written — and 9ea0197, main at the same moment.

[andrexibiza]

Fixed — title, body, and head-commit message now match the diff. Corrected:

• Title → refactor(main): extract shard s2 union (aux_config_cmd + model_picker) from main.py (matches the actual content of the extraction).
• Body → names hermes_cli/main.py (not run.py) and the two extracted modules (aux_config_cmd.py + model_picker.py), with the correct line math: 23 added / 859 deleted in main.py.
• Head commit → amended to the same corrected message (32ea28ba9b); tree is byte-identical to the previously reviewed tip 99e83f6de (verified: identical tree SHA), so nothing behavioral changed.

The kanban_db.py references were stale campaign-template metadata; hermes_cli/kanban_db.py is untouched, as the diff always showed. Validation: git diff --check clean; regression test tests/hermes_cli/test_w1_s2_extraction.py (shipped in this PR) covers the moved clusters.