docs(ai): handoff after CI parallelization + final test fix

Updates HANDOFF.md, CURRENT_TASK.md, and SESSION_LOG.md to reflect:
- PR #150 now contains the AI-provider test mock + caching + maxfail.
  Backend CI should be fully green for the first time in months.
- PR #151 stacked on #150: pytest-xdist with per-worker DBs. Local
  verification: 22m 27s → 4m 28s (5× speedup), 1076 passed both runs.
- DoD is now: merge #150, then #151, then add CI / backend
  (pull_request) to required status checks on main.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

.ai/HANDOFF.md

@@ -2,62 +2,75 @@
 
 # HANDOFF.md
 
-**Last updated:** 2026-04-25 06:12 EDT
+**Last updated:** 2026-04-25 (America/New_York)
 
-**Active task:** Restore green CI gate on `main` and lock it via branch protection. See [CURRENT_TASK.md](CURRENT_TASK.md).
+**Active task:** Land two stacked CI PRs (#150 + #151), then enable backend gate on `main`. See [CURRENT_TASK.md](CURRENT_TASK.md).
 
-**Branch:** `fix/ci-workflow-config`
+**Branch:** Currently on `fix/ci-workflow-config` (PR #150). The xdist work lives on `fix/ci-pytest-xdist` (PR #151), branched from #150.
 
-## Current state
+## Two open PRs to land in order
 
-Previous session fixed the 54 real backend failures left after #149. The default backend suite is now green locally:
+### PR #150 — `fix/ci-workflow-config` → main
 
-```bash
-docker exec resolutionflow_backend bash -lc 'pytest --override-ini="addopts=" -q > /tmp/full-backend.log 2>&1; code=$?; tail -n 160 /tmp/full-backend.log; exit $code'
-# 1076 passed, 35 deselected in 1347.41s (0:22:27)
-```
+Carries:
+- The Codex commit (`49f8856 wip(handoff): restore backend suite to green`) — fixes 54 backend test failures.
+- Workflow fixes: `DATABASE_TEST_URL` env, `actions/upload-artifact` v3 pin.
+- Most-recent commit (`e976fb4`):
+  - Mocks `_extract_template_parameters` in `test_record_decision_persists_and_bumps_state_version` (last test failing on CI; needed an AI provider key the runner doesn't have). Verified locally — passes.
+  - pip + npm caches in all three jobs.
+  - Drops `--cov-report=term-missing` (the custom "Display coverage summary" step prints the same info from JSON).
+  - Adds `--maxfail=10` so structural breakage fails fast.
 
-Targeted validation also passed:
+**Expected CI on this PR:** all three jobs green for the first time in months.
 
-- `tests/test_session_resolutions_api.py tests/test_session_sharing.py tests/test_session_suggested_fixes_api.py tests/test_survey.py tests/test_tenant_isolation_p0.py tests/test_tree_sharing.py tests/test_trees.py::TestTrees::test_delete_tree_cleans_up_folder_and_tag_assignments tests/test_uploads.py::test_delete_upload_forbidden_for_non_owner` → `73 passed`
-- PDF export tests → `3 passed`
-- Prompt/PSA/resolution/script-builder subset → `14 passed`
-- Admin/AI/branch subsets → `11 passed`
+### PR #151 — `fix/ci-pytest-xdist` → main (stacked on #150)
 
-## What changed
+Carries (on top of #150):
+- `pytest-xdist==3.6.1` in `requirements-dev.txt`.
+- `conftest.py` adds `_worker_db_url` + `_ensure_worker_db_exists`. Each xdist worker gets its own DB (`resolutionflow_test_gw0`, `gw1`, …) so the per-test `DROP SCHEMA public CASCADE` doesn't race across workers.
+- Workflow's pytest invocation gains `-n auto`.
 
-Production fixes:
-
-- CI/backend dev image now installs WeasyPrint system libraries.
-- Public share-token and survey routes are mounted outside tenant auth; protected share management remains tenant-protected.
-- Folder creation now persists `UserFolder.account_id`.
-- Script Builder save-to-library now persists `ScriptTemplate.account_id`.
-- Resolution output generation eager-loads `AISession.steps` to avoid async lazy-load `MissingGreenlet`.
-- AI session model now declares the generated `search_vector` column already present in Alembic, so `create_all` test schemas match runtime migrations.
-- Direct account-role update now rejects `"owner"`; ownership changes must use the transfer path.
-- Assistant prompt marker examples no longer include a literal executable `create_spin_off_ticket` payload.
-
-Test/harness fixes:
-
-- Test seeds updated for tenant-scoped `account_id` columns on sessions, branches, resolution outputs, script templates, PSA connections, folders, schedules, and categories.
-- Tests aligned with 404-not-403 resource-hiding policy.
-- Disabled-AI tests now restore both Anthropic and Google key settings.
-- Pytest harness closes pytest-asyncio's leftover clean loop and ignores known unclosed asyncio/asyncpg teardown ResourceWarnings that otherwise appear at arbitrary later setup points under `filterwarnings = error`.
+**Measured locally:** backend suite goes from `22m 27s` (serial, 1076 passed) → `4m 28s` (8 workers, 1076 passed). Same exit code, same test count.
 
 ## Immediate next steps
 
-1. Commit current working tree if not already committed with trailer:
-   `Co-Authored-By: Codex <noreply@openai.com>`.
-2. Check PR #150 status on Gitea. If both `CI / backend (pull_request)` and `CI / frontend (pull_request)` are green, merge it.
-3. After #150 merges, add `CI / backend (pull_request)` to required status checks on main:
+1. **Watch PR #150 CI** on its latest sha (`e976fb4`). Both `CI / backend (pull_request)` and `CI / frontend (pull_request)` should be green. Merge if so.
+2. **Watch PR #151 CI** after #150 merges. Once #151 is rebased / merged automatically, backend job time on subsequent runs should drop to the 4–6 min range.
+3. **Enable backend gate** on `main` branch protection — append `"CI / backend (pull_request)"` to `status_check_contexts`:
    ```bash
-   PATCH /repos/chihlasm/resolutionflow/branch_protections/main
-   { "status_check_contexts": ["CI / frontend (pull_request)", "CI / backend (pull_request)"] }
+   curl -X PATCH -H "Authorization: token $GITEA_TOKEN" \
+     "https://gitea.resolutionflow.com/api/v1/repos/chihlasm/resolutionflow/branch_protections/main" \
+     -H "Content-Type: application/json" \
+     -d '{"status_check_contexts": ["CI / frontend (pull_request)", "CI / backend (pull_request)"]}'
    ```
-   `$GITEA_TOKEN` is in `.claude/settings.local.json`.
-4. Run/confirm frontend lint if needed for the final DoD item (`npm run lint` was already green after #149, but this session did not rerun it).
+4. **Optional: also gate `CI / e2e (pull_request)`** once that job has run cleanly a few times. The artifact-v3 fix means it can finally run; we haven't verified its actual outcome yet.
+
+## Uncommitted state
+
+Working tree clean (after this handoff commit).
+
+## Branch protection on main (current)
+
+- PR-only merges
+- `CI / frontend (pull_request)` required
+- Force-push blocked
+- No review required (solo)
+
+## Recently merged on main
+
+- `f27f671` — PR #149: fix(ci): frontend lint to zero errors + test-DB schema fix + dev-deps installable
+- `06593a4` — PR #148: fix(tests): repair two pre-existing bugs blocking backend CI
+- `32fae2c` — PR #147: feat: FlowPilot migration — Phase 1-9 + Phase 9 bug fixes + QA fixture harness
+- `16060d2` — PR #141: feat: PSA ticket management
 
 ## Open questions
 
-- PR #150 was not rechecked or merged in this session.
-- Branch protection was not updated in this session.
+- One known concern with `--maxfail=10`: if a single bad commit produces 11+ legitimate failures, CI bails before reporting them all. Acceptable trade-off — the alternative is burning 25 min on a structural break.
+- pytest-xdist load distribution is the default file-scoped balance. If one worker consistently gets the slow tests, switch to `--dist worksteal` (xdist 3.x). Not worth tuning preemptively.
+
+## Useful breadcrumbs
+
+- `backend/scripts/seed_phase9_qa_fixtures.py` pre-bakes Phase 9 QA fixtures.
+- `.gstack/qa-reports/phase9-20260424-232700/REPORT.md` — full QA report from the FlowPilot session.
+- gstack is in team mode for this repo. `/browse` Chromium needs `CONTAINER=1` env (see `~/.claude/skills/gstack/browse/src/browser-manager.ts:188`).
+- Per-worker test DBs accumulate on the postgres service. Cheap to leave around; cleanup if it ever bothers anyone.