fix(honcho): restore profile fallback and backfill memory files

[stephenschoettler]

Bug Description

honcho_profile could return an empty profile even when Honcho session context already had a populated peer card. Existing sessions also never imported USER.md, MEMORY.md, or SOUL.md if they were created before the migration path ran.

Root Cause

• get_peer_card() only trusted the direct peer-card lookup path.
• Existing-session migration only ran for brand new Honcho sessions.

Fix

• fall back to session context peer-card data when direct peer-card lookup is empty
• make honcho_profile fall back to session context before returning an empty result
• backfill memory files into existing non-per-session Honcho sessions once per content hash and persist a local backfill marker
• add regression coverage for both fallback and backfill paths

How to Verify

1. Run source venv/bin/activate && pytest tests/honcho_plugin/test_session.py -q.
2. Confirm honcho_profile returns session-context facts when direct peer-card lookup is empty.
3. Confirm an existing session backfills memory files only once for a given memory-file hash.

Test Plan

• Added regression test for this bug
• Existing tests still pass
• Manual verification of the fix

Manual verification:

• source venv/bin/activate && pytest tests/honcho_plugin/test_session.py -q
• source venv/bin/activate && python - <<'PY' from plugins.memory.honcho.client import HonchoClientConfig, get_honcho_client from plugins.memory.honcho.session import HonchoSessionManager cfg = HonchoClientConfig.from_global_config() client = get_honcho_client(cfg) mgr = HonchoSessionManager(honcho=client, config=cfg) s = mgr.get_or_create(cfg.resolve_session_name(cwd='/home/w0lf')) print(mgr.get_peer_card(s.key, peer='user')[:8]) PY

Risk Assessment

Low - the fallback only activates when the direct card lookup is empty, and the backfill path is limited to non-per-session sessions with local memory files and a new content hash.

[stephenschoettler]

Superseded by #13369. The two failing checks here were caused by this branch being based on a stale main snapshot, which pulled in unrelated failures and triggered the supply-chain scan on old setup.py changes. I rebuilt the Honcho patch cleanly on current main and opened a fresh PR with only the 3 intended files changed.

[stephenschoettler]

Superseded again by #13375. #13369 removed the unrelated test failures but was still based on a stale main snapshot, so the supply-chain audit kept seeing unrelated setup.py changes. #13375 is rebased onto current main and contains only the 3 intended Honcho files.

[stephenschoettler]

Closing as superseded by #13375.