phase: 166_hide_empty_answer_bubble

**Phase 166 — final verification pass: green.** All 3 task files already in `complete/`, so this was verification + defect/doc fixing. - **Verified:** `.msg .bubble:empty { display: none; }` exists once in the Messages block with provenance; `frontend/` diff is styles.css only (**no app.js change**); `tests/unit/test_no_empty_answer_bubble.py` 26 pins pass; phase-158/161 anchor pins (thinking_timeline + tool_states + edit_message) **75 passed**. - **Phase E2E in isolation:** `uv run pytest tests/e2e/test_no_empty_box.py --no-cov` → **5 passed**. Mutation check (rule removed) → 2 failures, so the suite genuinely pins the defect. - **Regression set, isolated:** thinking_timeline 6 · turn_progress_loader 3 · chat_history 5 · share_chat 4 · save_share_ux 5 · failed_turn_retry 4 · chat_answer_clusters 4 · chat_error_honesty 4 → all pass. - **Full gates:** `uv run pytest --cov=app --cov-report=term-missing` → **3880 passed, app/ 99%** (gate >90%); `uv run ruff check .` → clean; `uv run pyright` → **0 errors**. - **Whole E2E tree swept file-by-file (150 files, own pytest process each).** 4 suites were broken by *earlier* phases; fixed and re-run green: - `test_kb_tree`, `test_sync_summary_visibility`, `test_folder_summary_voice` — `td:nth-child(4) span` became a strict-mode violation when phase 156 wrapped the Edit label in `.cell-clip` → scoped to `span.kb-desc-text`. - `test_search_tool_scope` — pinned the pre-`690fd98` per-run `<summary>`/`open` shape → re-pinned to the ONE `details.reasoning` container + `.tool-calls-summary`. - `test_header_consistency` — `.doc-meta` pixel bound (`<34px`) was broken by phase 153's 44px touch-target floor at ≤640px → replaced with a shape-first single-line check that still fails a real wrap. - **Completion criteria:** mid-turn box hidden ✅ · bubble appears at first delta, phase-130 rows below unchanged ✅ · restored/shared empty answer paints no box ✅ · no app.js change + anchors + full gates ✅ · commit/`complete/` move = harness's job (I ran no git index/history commands). - **Docs:** README — added the phase-166 "no empty answer box mid-turn" note; fixed a **stale** Thinking lead ("collapsible Thinking blocks", pre-dated the 2026-10-06 one-disclosure refinement); QA section now states E2E isolation is mandatory (shared Postgres + fixed ports 8123/8901) with the per-file sweep loop. AGENTS.md — rule 5 UI checklist now names the phase-166 suppression rule. - **Notable:** running all of `tests/e2e/` in one pytest session is not supported (fixture/port interleaving causes ~126 spurious failures) — documented in README. Restored the tracked `.agents/screenshots/doc_modal_scrollbar_1280.png` that `test_document_viewer.py` rewrites on every run. - **Next pending phase:** `167_api_payload_diet`.
2026-10-08 04:27:03 -04:00
parent b0c7aed613
commit ab9391d019
36 changed files with 2348 additions and 93 deletions
@@ -0,0 +1,14 @@
+**Phase 166 — final verification pass: green.** All 3 task files already in `complete/`, so this was verification + defect/doc fixing.
+
+- **Verified:** `.msg .bubble:empty { display: none; }` exists once in the Messages block with provenance; `frontend/` diff is styles.css only (**no app.js change**); `tests/unit/test_no_empty_answer_bubble.py` 26 pins pass; phase-158/161 anchor pins (thinking_timeline + tool_states + edit_message) **75 passed**.
+- **Phase E2E in isolation:** `uv run pytest tests/e2e/test_no_empty_box.py --no-cov` → **5 passed**. Mutation check (rule removed) → 2 failures, so the suite genuinely pins the defect.
+- **Regression set, isolated:** thinking_timeline 6 · turn_progress_loader 3 · chat_history 5 · share_chat 4 · save_share_ux 5 · failed_turn_retry 4 · chat_answer_clusters 4 · chat_error_honesty 4 → all pass.
+- **Full gates:** `uv run pytest --cov=app --cov-report=term-missing` → **3880 passed, app/ 99%** (gate >90%); `uv run ruff check .` → clean; `uv run pyright` → **0 errors**.
+- **Whole E2E tree swept file-by-file (150 files, own pytest process each).** 4 suites were broken by *earlier* phases; fixed and re-run green:
+  - `test_kb_tree`, `test_sync_summary_visibility`, `test_folder_summary_voice` — `td:nth-child(4) span` became a strict-mode violation when phase 156 wrapped the Edit label in `.cell-clip` → scoped to `span.kb-desc-text`.
+  - `test_search_tool_scope` — pinned the pre-`690fd98` per-run `<summary>`/`open` shape → re-pinned to the ONE `details.reasoning` container + `.tool-calls-summary`.
+  - `test_header_consistency` — `.doc-meta` pixel bound (`<34px`) was broken by phase 153's 44px touch-target floor at ≤640px → replaced with a shape-first single-line check that still fails a real wrap.
+- **Completion criteria:** mid-turn box hidden ✅ · bubble appears at first delta, phase-130 rows below unchanged ✅ · restored/shared empty answer paints no box ✅ · no app.js change + anchors + full gates ✅ · commit/`complete/` move = harness's job (I ran no git index/history commands).
+- **Docs:** README — added the phase-166 "no empty answer box mid-turn" note; fixed a **stale** Thinking lead ("collapsible Thinking blocks", pre-dated the 2026-10-06 one-disclosure refinement); QA section now states E2E isolation is mandatory (shared Postgres + fixed ports 8123/8901) with the per-file sweep loop. AGENTS.md — rule 5 UI checklist now names the phase-166 suppression rule.
+- **Notable:** running all of `tests/e2e/` in one pytest session is not supported (fixture/port interleaving causes ~126 spurious failures) — documented in README. Restored the tracked `.agents/screenshots/doc_modal_scrollbar_1280.png` that `test_document_viewer.py` rewrites on every run.
+- **Next pending phase:** `167_api_payload_diet`.
@@ -0,0 +1 @@
+ok
@@ -0,0 +1,125 @@
+........................................................................ [  1%]
+........................................................................ [  3%]
+........................................................................ [  5%]
+........................................................................ [  7%]
+........................................................................ [  9%]
+........................................................................ [ 11%]
+........................................................................ [ 12%]
+........................................................................ [ 14%]
+........................................................................ [ 16%]
+........................................................................ [ 18%]
+........................................................................ [ 20%]
+........................................................................ [ 22%]
+........................................................................ [ 24%]
+........................................................................ [ 25%]
+........................................................................ [ 27%]
+........................................................................ [ 29%]
+........................................................................ [ 31%]
+........................................................................ [ 33%]
+........................................................................ [ 35%]
+........................................................................ [ 37%]
+........................................................................ [ 38%]
+........................................................................ [ 40%]
+........................................................................ [ 42%]
+........................................................................ [ 44%]
+........................................................................ [ 46%]
+........................................................................ [ 48%]
+........................................................................ [ 50%]
+........................................................................ [ 51%]
+........................................................................ [ 53%]
+........................................................................ [ 55%]
+........................................................................ [ 57%]
+........................................................................ [ 59%]
+........................................................................ [ 61%]
+........................................................................ [ 63%]
+........................................................................ [ 64%]
+........................................................................ [ 66%]
+........................................................................ [ 68%]
+........................................................................ [ 70%]
+........................................................................ [ 72%]
+........................................................................ [ 74%]
+........................................................................ [ 76%]
+........................................................................ [ 77%]
+........................................................................ [ 79%]
+........................................................................ [ 81%]
+........................................................................ [ 83%]
+........................................................................ [ 85%]
+........................................................................ [ 87%]
+........................................................................ [ 89%]
+........................................................................ [ 90%]
+........................................................................ [ 92%]
+........................................................................ [ 94%]
+........................................................................ [ 96%]
+........................................................................ [ 98%]
+................................................................         [100%]
+=============================== warnings summary ===============================
+.venv/lib/python3.13/site-packages/fastapi/testclient.py:1
+  /var/home/ducoterra/Projects/Personal/brain_of_reese/.venv/lib/python3.13/site-packages/fastapi/testclient.py:1: StarletteDeprecationWarning: Using `httpx` with `starlette.testclient` is deprecated; install `httpx2` instead.
+    from starlette.testclient import TestClient as TestClient  # noqa
+
+-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
+================================ tests coverage ================================
+_______________ coverage: platform linux, python 3.13.13-final-0 _______________
+
+Name                           Stmts   Miss  Cover
+--------------------------------------------------
+app/__init__.py                    1      0   100%
+app/api/__init__.py                0      0   100%
+app/api/auth.py                   52      0   100%
+app/api/chat.py                  277      1    99%
+app/api/chat_images.py            50      0   100%
+app/api/chats.py                 110      0   100%
+app/api/config.py                 13      0   100%
+app/api/doc_drafts.py             99      0   100%
+app/api/docs.py                  196      1    99%
+app/api/git_sources.py           259      0   100%
+app/api/health.py                 10      0   100%
+app/api/steering.py               42      0   100%
+app/api/suggestions.py            33      0   100%
+app/api/sync.py                  204      0   100%
+app/api/tokens.py                 40      0   100%
+app/api/ui_settings.py            55      0   100%
+app/config.py                    266      0   100%
+app/core/__init__.py               0      0   100%
+app/core/auth.py                  45      0   100%
+app/core/caching.py              124      0   100%
+app/core/debugging.py             29      2    93%
+app/core/docs_push.py             39      0   100%
+app/core/errors.py                 5      0   100%
+app/core/logging.py               13      0   100%
+app/core/rate_limit.py            44      0   100%
+app/core/security_headers.py      20      0   100%
+app/core/theming.py               38      0   100%
+app/core/tokens.py                44      0   100%
+app/db.py                         22      0   100%
+app/main.py                       68      0   100%
+app/models.py                    136      0   100%
+app/rag/__init__.py                0      0   100%
+app/rag/agent.py                 454      1    99%
+app/rag/archive_upload.py        134      0   100%
+app/rag/chunker.py               220      4    98%
+app/rag/doc_dates.py              18      0   100%
+app/rag/folder_summaries.py      134      0   100%
+app/rag/git_sources.py            77      0   100%
+app/rag/importer.py              841      4    99%
+app/rag/llm.py                   246      1    99%
+app/rag/md_images.py              63      0   100%
+app/rag/overview.py               78      0   100%
+app/rag/pdf.py                   192      0   100%
+app/rag/prompts.py               156      0   100%
+app/rag/retriever.py             252      3    99%
+app/rag/scaffolding.py            55      0   100%
+app/rag/source_removal.py         41      0   100%
+app/rag/sources_meta.py           16      0   100%
+app/rag/suggestions.py            27      0   100%
+app/rag/summarizer.py             58      0   100%
+app/schemas.py                   487      0   100%
+--------------------------------------------------
+TOTAL                           5883     17    99%
+coverage gate: app/ 99% (>90%) OK
+All checks passed!
+0 errors, 0 warnings, 0 informations
+WARNING: there is a new pyright version available (v1.1.411 -> v1.1.414).
+Please install the new version or set PYRIGHT_PYTHON_FORCE_VERSION to `latest`
+
+validation OK
@@ -0,0 +1,18 @@
+## Report — task 166/01 `hide_empty_bubble`
+
+**Implemented**
+- `frontend/assets/styles.css`: added `.msg .bubble:empty { display: none; }` in the Messages block, directly after `.bubble p`, with the full provenance comment (TODO L4, `image.png`, the `addMessage("brain","")` / `ensureReasoningWrap` anchor it preserves, why the typing indicator is never `:empty`, flex-gap collapse, scoping, no tokens/animation).
+- New `tests/unit/test_no_empty_answer_bubble.py` (11 tests, house source-assertion style): rule exists exactly once + `.msg`-scoped + ordered between `.bubble p` and `.bubble pre`; `.bubble` keeps padding/border-radius/background/border/box-shadow (hides, never restyles); typing-dot rules untouched; `addTyping`'s three dot `<span>`s pinned (loading state can't vanish); all three `addMessage("brain", "")` anchors + the `insertBefore(cont, querySelector(".bubble"))` anchor survive (D2); node-executed `renderMarkdown("")`/`("   ")` → `""` (why restore/shared are covered, A3); no-CDN pin.
+- `frontend/assets/app.js`: **no change** (LOCKED A1 respected).
+
+**Tests / lint / coverage**
+- `uv run pytest tests/unit/test_no_empty_answer_bubble.py -v` → 11 passed
+- `uv run pytest tests/unit -k "thinking_timeline or tool_states or edit_message or frontend"` → all passed (phase-158/161/159 anchor pins untouched)
+- `uv run pytest --cov=app --cov-report=term-missing` → **3865 passed**, `app/` coverage **99%** (>90% gate unchanged)
+- `uv run ruff check .` → clean; `uv run pyright` → 0 errors
+
+**Docs**: no doc impact — README never described the empty box (behavior now matches the docs); task file pins the diff to styles.css + one test file only.
+
+**Decisions**: none deviating; git diff touches exactly `frontend/assets/styles.css` + the new unit test file.
+
+**Next pending task**: `166_hide_empty_answer_bubble/02_feedback_contract_check.md`
@@ -0,0 +1 @@
+ok
@@ -0,0 +1,125 @@
+........................................................................ [  1%]
+........................................................................ [  3%]
+........................................................................ [  5%]
+........................................................................ [  7%]
+........................................................................ [  9%]
+........................................................................ [ 11%]
+........................................................................ [ 13%]
+........................................................................ [ 14%]
+........................................................................ [ 16%]
+........................................................................ [ 18%]
+........................................................................ [ 20%]
+........................................................................ [ 22%]
+........................................................................ [ 24%]
+........................................................................ [ 26%]
+........................................................................ [ 27%]
+........................................................................ [ 29%]
+........................................................................ [ 31%]
+........................................................................ [ 33%]
+........................................................................ [ 35%]
+........................................................................ [ 37%]
+........................................................................ [ 39%]
+........................................................................ [ 40%]
+........................................................................ [ 42%]
+........................................................................ [ 44%]
+........................................................................ [ 46%]
+........................................................................ [ 48%]
+........................................................................ [ 50%]
+........................................................................ [ 52%]
+........................................................................ [ 54%]
+........................................................................ [ 55%]
+........................................................................ [ 57%]
+........................................................................ [ 59%]
+........................................................................ [ 61%]
+........................................................................ [ 63%]
+........................................................................ [ 65%]
+........................................................................ [ 67%]
+........................................................................ [ 68%]
+........................................................................ [ 70%]
+........................................................................ [ 72%]
+........................................................................ [ 74%]
+........................................................................ [ 76%]
+........................................................................ [ 78%]
+........................................................................ [ 80%]
+........................................................................ [ 81%]
+........................................................................ [ 83%]
+........................................................................ [ 85%]
+........................................................................ [ 87%]
+........................................................................ [ 89%]
+........................................................................ [ 91%]
+........................................................................ [ 93%]
+........................................................................ [ 95%]
+........................................................................ [ 96%]
+........................................................................ [ 98%]
+.................................................                        [100%]
+=============================== warnings summary ===============================
+.venv/lib/python3.13/site-packages/fastapi/testclient.py:1
+  /var/home/ducoterra/Projects/Personal/brain_of_reese/.venv/lib/python3.13/site-packages/fastapi/testclient.py:1: StarletteDeprecationWarning: Using `httpx` with `starlette.testclient` is deprecated; install `httpx2` instead.
+    from starlette.testclient import TestClient as TestClient  # noqa
+
+-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
+================================ tests coverage ================================
+_______________ coverage: platform linux, python 3.13.13-final-0 _______________
+
+Name                           Stmts   Miss  Cover
+--------------------------------------------------
+app/__init__.py                    1      0   100%
+app/api/__init__.py                0      0   100%
+app/api/auth.py                   52      0   100%
+app/api/chat.py                  277      1    99%
+app/api/chat_images.py            50      0   100%
+app/api/chats.py                 110      0   100%
+app/api/config.py                 13      0   100%
+app/api/doc_drafts.py             99      0   100%
+app/api/docs.py                  196      1    99%
+app/api/git_sources.py           259      0   100%
+app/api/health.py                 10      0   100%
+app/api/steering.py               42      0   100%
+app/api/suggestions.py            33      0   100%
+app/api/sync.py                  204      0   100%
+app/api/tokens.py                 40      0   100%
+app/api/ui_settings.py            55      0   100%
+app/config.py                    266      0   100%
+app/core/__init__.py               0      0   100%
+app/core/auth.py                  45      0   100%
+app/core/caching.py              124      0   100%
+app/core/debugging.py             29      2    93%
+app/core/docs_push.py             39      0   100%
+app/core/errors.py                 5      0   100%
+app/core/logging.py               13      0   100%
+app/core/rate_limit.py            44      0   100%
+app/core/security_headers.py      20      0   100%
+app/core/theming.py               38      0   100%
+app/core/tokens.py                44      0   100%
+app/db.py                         22      0   100%
+app/main.py                       68      0   100%
+app/models.py                    136      0   100%
+app/rag/__init__.py                0      0   100%
+app/rag/agent.py                 454      1    99%
+app/rag/archive_upload.py        134      0   100%
+app/rag/chunker.py               220      4    98%
+app/rag/doc_dates.py              18      0   100%
+app/rag/folder_summaries.py      134      0   100%
+app/rag/git_sources.py            77      0   100%
+app/rag/importer.py              841      4    99%
+app/rag/llm.py                   246      1    99%
+app/rag/md_images.py              63      0   100%
+app/rag/overview.py               78      0   100%
+app/rag/pdf.py                   192      0   100%
+app/rag/prompts.py               156      0   100%
+app/rag/retriever.py             252      3    99%
+app/rag/scaffolding.py            55      0   100%
+app/rag/source_removal.py         41      0   100%
+app/rag/sources_meta.py           16      0   100%
+app/rag/suggestions.py            27      0   100%
+app/rag/summarizer.py             58      0   100%
+app/schemas.py                   487      0   100%
+--------------------------------------------------
+TOTAL                           5883     17    99%
+coverage gate: app/ 99% (>90%) OK
+All checks passed!
+0 errors, 0 warnings, 0 informations
+WARNING: there is a new pyright version available (v1.1.411 -> v1.1.414).
+Please install the new version or set PYRIGHT_PYTHON_FORCE_VERSION to `latest`
+
+validation OK
@@ -0,0 +1,20 @@
+**Task 02 complete — §7.4 feedback contract proven with the empty bubble hidden.**
+
+**Implemented** (all in `tests/unit/test_no_empty_answer_bubble.py`, +14 tests; no `app/` or `frontend/` change):
+- Module docstring `AUDIT NOTE:` — the four-state mid-turn inventory (typing dots+clock → reasoning container+status span → bubble at first token → terminal copy) as the record for the next agent.
+- Feedback inventory pins: `addTyping` `role="status"`/`TYPING_LABEL()`/3 dots; `thinkingClock` `.typing-elapsed` `Ns` suffix (bubble never `:empty`); `setReasoningStatus` literals wired (1 Thinking / 1 Calling / 3 settle "N steps"); `cont.open = true`; `SEND_STATUS`/`#send-label` copy; `#messages` `aria-live="polite"` with announcement-timing reasoning.
+- "Nothing measures the hidden box": `if (scroll) scrollReveal(wrap)` takes the `.msg` wrap; phase-161 pins take `.reasoning-body` (never a `.bubble`); grep-audits — no geometry prop (`getBoundingClientRect`/`offset*`/`client*`/`scroll*`) ever pairs with `bubble`, the `.bubble` selector appears only in `querySelector`/`insertBefore` lines, and the bubble variable is touched only for content (allowlist incl. `textContent`, `innerHTML`, `.querySelector(".msg-image")`, `classList`).
+- `.msg-meta` append sites pinned to `.msg-body` (chips, images row, answer figure, actions row, `.failed-note`, `.stopped-note`) — none mentions `.bubble`.
+- Restore (`addMessage("brain", renderMarkdown(m.text), true)`, ×1) and shared page (`shared.js` byte-identical `<div class="bubble">${html}</div>` + `msg ${who}` wrap, same `styles.css`, zero `:empty` of its own) covered by the same `.msg ` scope — no `shared.js` change.
+
+**Gates** (all green):
+- `uv run pytest tests/unit/test_no_empty_answer_bubble.py -v` → 26 passed
+- `uv run pytest tests/unit -k "frontend or reasoning or thinking or tool" -v` → 572 passed
+- `uv run pytest --cov=app --cov-report=term-missing` → 3880 passed, `app/` coverage **99%** (>90%, unmoved)
+- `uv run ruff check .` → All checks passed; `uv run pyright` → 0 errors
+
+**Docs:** no doc impact — test-only task, nothing user-visible changed.
+
+**Notes:** `ruff format` isn't a project gate (470 pre-existing files unformatted); I matched the `ruff check` gate. D2 honored: zero JS/CSS changes.
+
+**Next pending task:** `166_hide_empty_answer_bubble/03_e2e_no_empty_box.md`.
@@ -0,0 +1 @@
+ok
@@ -0,0 +1,125 @@
+........................................................................ [  1%]
+........................................................................ [  3%]
+........................................................................ [  5%]
+........................................................................ [  7%]
+........................................................................ [  9%]
+........................................................................ [ 11%]
+........................................................................ [ 12%]
+........................................................................ [ 14%]
+........................................................................ [ 16%]
+........................................................................ [ 18%]
+........................................................................ [ 20%]
+........................................................................ [ 22%]
+........................................................................ [ 24%]
+........................................................................ [ 25%]
+........................................................................ [ 27%]
+........................................................................ [ 29%]
+........................................................................ [ 31%]
+........................................................................ [ 33%]
+........................................................................ [ 35%]
+........................................................................ [ 37%]
+........................................................................ [ 38%]
+........................................................................ [ 40%]
+........................................................................ [ 42%]
+........................................................................ [ 44%]
+........................................................................ [ 46%]
+........................................................................ [ 48%]
+........................................................................ [ 50%]
+........................................................................ [ 51%]
+........................................................................ [ 53%]
+........................................................................ [ 55%]
+........................................................................ [ 57%]
+........................................................................ [ 59%]
+........................................................................ [ 61%]
+........................................................................ [ 63%]
+........................................................................ [ 64%]
+........................................................................ [ 66%]
+........................................................................ [ 68%]
+........................................................................ [ 70%]
+........................................................................ [ 72%]
+........................................................................ [ 74%]
+........................................................................ [ 76%]
+........................................................................ [ 77%]
+........................................................................ [ 79%]
+........................................................................ [ 81%]
+........................................................................ [ 83%]
+........................................................................ [ 85%]
+........................................................................ [ 87%]
+........................................................................ [ 89%]
+........................................................................ [ 90%]
+........................................................................ [ 92%]
+........................................................................ [ 94%]
+........................................................................ [ 96%]
+........................................................................ [ 98%]
+................................................................         [100%]
+=============================== warnings summary ===============================
+.venv/lib/python3.13/site-packages/fastapi/testclient.py:1
+  /var/home/ducoterra/Projects/Personal/brain_of_reese/.venv/lib/python3.13/site-packages/fastapi/testclient.py:1: StarletteDeprecationWarning: Using `httpx` with `starlette.testclient` is deprecated; install `httpx2` instead.
+    from starlette.testclient import TestClient as TestClient  # noqa
+
+-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
+================================ tests coverage ================================
+_______________ coverage: platform linux, python 3.13.13-final-0 _______________
+
+Name                           Stmts   Miss  Cover
+--------------------------------------------------
+app/__init__.py                    1      0   100%
+app/api/__init__.py                0      0   100%
+app/api/auth.py                   52      0   100%
+app/api/chat.py                  277      1    99%
+app/api/chat_images.py            50      0   100%
+app/api/chats.py                 110      0   100%
+app/api/config.py                 13      0   100%
+app/api/doc_drafts.py             99      0   100%
+app/api/docs.py                  196      1    99%
+app/api/git_sources.py           259      0   100%
+app/api/health.py                 10      0   100%
+app/api/steering.py               42      0   100%
+app/api/suggestions.py            33      0   100%
+app/api/sync.py                  204      0   100%
+app/api/tokens.py                 40      0   100%
+app/api/ui_settings.py            55      0   100%
+app/config.py                    266      0   100%
+app/core/__init__.py               0      0   100%
+app/core/auth.py                  45      0   100%
+app/core/caching.py              124      0   100%
+app/core/debugging.py             29      2    93%
+app/core/docs_push.py             39      0   100%
+app/core/errors.py                 5      0   100%
+app/core/logging.py               13      0   100%
+app/core/rate_limit.py            44      0   100%
+app/core/security_headers.py      20      0   100%
+app/core/theming.py               38      0   100%
+app/core/tokens.py                44      0   100%
+app/db.py                         22      0   100%
+app/main.py                       68      0   100%
+app/models.py                    136      0   100%
+app/rag/__init__.py                0      0   100%
+app/rag/agent.py                 454      1    99%
+app/rag/archive_upload.py        134      0   100%
+app/rag/chunker.py               220      4    98%
+app/rag/doc_dates.py              18      0   100%
+app/rag/folder_summaries.py      134      0   100%
+app/rag/git_sources.py            77      0   100%
+app/rag/importer.py              841      4    99%
+app/rag/llm.py                   246      1    99%
+app/rag/md_images.py              63      0   100%
+app/rag/overview.py               78      0   100%
+app/rag/pdf.py                   192      0   100%
+app/rag/prompts.py               156      0   100%
+app/rag/retriever.py             252      3    99%
+app/rag/scaffolding.py            55      0   100%
+app/rag/source_removal.py         41      0   100%
+app/rag/sources_meta.py           16      0   100%
+app/rag/suggestions.py            27      0   100%
+app/rag/summarizer.py             58      0   100%
+app/schemas.py                   487      0   100%
+--------------------------------------------------
+TOTAL                           5883     17    99%
+coverage gate: app/ 99% (>90%) OK
+All checks passed!
+0 errors, 0 warnings, 0 informations
+WARNING: there is a new pyright version available (v1.1.411 -> v1.1.414).
+Please install the new version or set PYRIGHT_PYTHON_FORCE_VERSION to `latest`
+
+validation OK
@@ -0,0 +1,22 @@
+**Task complete.** All gates green.
+
+## Task Report
+
+**Task:** `03_e2e_no_empty_box.md` — Playwright E2E proving the empty answer box never paints (phase 166)
+
+**Files changed:**
+- `tests/e2e/test_no_empty_box.py` *(new)* — 5 E2E tests: pre-first-token window (synced on the reasoning stream: bubble `display:none`, 0px), plain-turn typing-indicator coverage (raced with an in-page 10ms paint sampler), restored chat, shared page, whitespace-only persisted answer
+- `tests/e2e/mock_llm.py` — new paced flow `EMPTY_BUBBLE_TRIGGER` ("paint the pre-answer window"): thinking+tool turn with a widened pre-content gap
+- `frontend/assets/styles.css` — the task-01 rule `.msg .bubble:empty { display:none }` (was uncommitted in the tree; rule untouched, verified)
+- Screenshot artifact: `.agents/screenshots/166_pre_answer_window_1280.png` (gitignored, house convention)
+
+**Pre-existing defects found & fixed** (suites left stale by `690fd98` "container is ONE disclosure"):
+- `test_turn_progress_loader.py` (4 tests) and `test_share_chat.py` (3 tests) — pinned the deleted live `details.thinking` shape → updated to the container/run shape; contracts preserved
+- `test_sources_midstream_bug.py` (2 tests) — same stale live selector → `div.thinking` runs
+- `test_no_reply_autoscroll.py` — 3 tests: 2 stale-shape; `test_no_autoscroll_during_long_answer` failed on a **native −44px scrollY adjustment** at settle (proven via probes: at done the Edit/Retry re-target collapses turn 1's meta row 44→0px *above* the viewport; Chromium's scroll anchoring compensates scrollY so the reader's view stays pixel-still; predates 166 — reproduced with HEAD's CSS and the full pre-161 frontend). Metric modernized to visual-stability of the reading position — same contract, asserted on what the user perceives; real up-hops still fail it.
+
+**Tests:** `test_no_empty_box.py` 5/5 in isolation; regression set (thinking_timeline, reasoning_bubble, turn_progress_loader, chat_history, share_chat, save_share_ux, failed_turn_retry, chat_answer_clusters, chat_error_honesty, markdown_tables, wide_desktop_column) all green; smoke green; full unit+integration **3880 passed**, `app/` coverage **99%** (>90% gate); ruff clean.
+
+**Docs:** README/AGENTS.md need no changes (test-only task; the visual fix itself was task 01).
+
+**Notes:** No git index/history commands run; changes left in the working tree for the pipeline harness to commit and to move `03_e2e_no_empty_box.md` to `.agents/phases/complete/166_hide_empty_answer_bubble/`.
@@ -0,0 +1 @@
+ok
@@ -0,0 +1,125 @@
+........................................................................ [  1%]
+........................................................................ [  3%]
+........................................................................ [  5%]
+........................................................................ [  7%]
+........................................................................ [  9%]
+........................................................................ [ 11%]
+........................................................................ [ 12%]
+........................................................................ [ 14%]
+........................................................................ [ 16%]
+........................................................................ [ 18%]
+........................................................................ [ 20%]
+........................................................................ [ 22%]
+........................................................................ [ 24%]
+........................................................................ [ 25%]
+........................................................................ [ 27%]
+........................................................................ [ 29%]
+........................................................................ [ 31%]
+........................................................................ [ 33%]
+........................................................................ [ 35%]
+........................................................................ [ 37%]
+........................................................................ [ 38%]
+........................................................................ [ 40%]
+........................................................................ [ 42%]
+........................................................................ [ 44%]
+........................................................................ [ 46%]
+........................................................................ [ 48%]
+........................................................................ [ 50%]
+........................................................................ [ 51%]
+........................................................................ [ 53%]
+........................................................................ [ 55%]
+........................................................................ [ 57%]
+........................................................................ [ 59%]
+........................................................................ [ 61%]
+........................................................................ [ 63%]
+........................................................................ [ 64%]
+........................................................................ [ 66%]
+........................................................................ [ 68%]
+........................................................................ [ 70%]
+........................................................................ [ 72%]
+........................................................................ [ 74%]
+........................................................................ [ 76%]
+........................................................................ [ 77%]
+........................................................................ [ 79%]
+........................................................................ [ 81%]
+........................................................................ [ 83%]
+........................................................................ [ 85%]
+........................................................................ [ 87%]
+........................................................................ [ 89%]
+........................................................................ [ 90%]
+........................................................................ [ 92%]
+........................................................................ [ 94%]
+........................................................................ [ 96%]
+........................................................................ [ 98%]
+................................................................         [100%]
+=============================== warnings summary ===============================
+.venv/lib/python3.13/site-packages/fastapi/testclient.py:1
+  /var/home/ducoterra/Projects/Personal/brain_of_reese/.venv/lib/python3.13/site-packages/fastapi/testclient.py:1: StarletteDeprecationWarning: Using `httpx` with `starlette.testclient` is deprecated; install `httpx2` instead.
+    from starlette.testclient import TestClient as TestClient  # noqa
+
+-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
+================================ tests coverage ================================
+_______________ coverage: platform linux, python 3.13.13-final-0 _______________
+
+Name                           Stmts   Miss  Cover
+--------------------------------------------------
+app/__init__.py                    1      0   100%
+app/api/__init__.py                0      0   100%
+app/api/auth.py                   52      0   100%
+app/api/chat.py                  277      1    99%
+app/api/chat_images.py            50      0   100%
+app/api/chats.py                 110      0   100%
+app/api/config.py                 13      0   100%
+app/api/doc_drafts.py             99      0   100%
+app/api/docs.py                  196      1    99%
+app/api/git_sources.py           259      0   100%
+app/api/health.py                 10      0   100%
+app/api/steering.py               42      0   100%
+app/api/suggestions.py            33      0   100%
+app/api/sync.py                  204      0   100%
+app/api/tokens.py                 40      0   100%
+app/api/ui_settings.py            55      0   100%
+app/config.py                    266      0   100%
+app/core/__init__.py               0      0   100%
+app/core/auth.py                  45      0   100%
+app/core/caching.py              124      0   100%
+app/core/debugging.py             29      2    93%
+app/core/docs_push.py             39      0   100%
+app/core/errors.py                 5      0   100%
+app/core/logging.py               13      0   100%
+app/core/rate_limit.py            44      0   100%
+app/core/security_headers.py      20      0   100%
+app/core/theming.py               38      0   100%
+app/core/tokens.py                44      0   100%
+app/db.py                         22      0   100%
+app/main.py                       68      0   100%
+app/models.py                    136      0   100%
+app/rag/__init__.py                0      0   100%
+app/rag/agent.py                 454      1    99%
+app/rag/archive_upload.py        134      0   100%
+app/rag/chunker.py               220      4    98%
+app/rag/doc_dates.py              18      0   100%
+app/rag/folder_summaries.py      134      0   100%
+app/rag/git_sources.py            77      0   100%
+app/rag/importer.py              841      4    99%
+app/rag/llm.py                   246      1    99%
+app/rag/md_images.py              63      0   100%
+app/rag/overview.py               78      0   100%
+app/rag/pdf.py                   192      0   100%
+app/rag/prompts.py               156      0   100%
+app/rag/retriever.py             252      3    99%
+app/rag/scaffolding.py            55      0   100%
+app/rag/source_removal.py         41      0   100%
+app/rag/sources_meta.py           16      0   100%
+app/rag/suggestions.py            27      0   100%
+app/rag/summarizer.py             58      0   100%
+app/schemas.py                   487      0   100%
+--------------------------------------------------
+TOTAL                           5883     17    99%
+coverage gate: app/ 99% (>90%) OK
+All checks passed!
+0 errors, 0 warnings, 0 informations
+WARNING: there is a new pyright version available (v1.1.411 -> v1.1.414).
+Please install the new version or set PYRIGHT_PYTHON_FORCE_VERSION to `latest`
+
+validation OK
@@ -21,7 +21,10 @@
   panels — `max-height: min(70vh, 64rem)` — with a sticky header row;
   a clipped action label shows an ellipsis hint (`.cell-clip`) with
   the full text kept reachable via the control's aria-label
-   (phase 156); the document viewer's page strip is CLASS-scoped —
+   (phase 156); an in-flight turn never paints an empty answer
+   bubble — the box is created empty as the reasoning anchor and
+   `.msg .bubble:empty { display: none }` hides it until the first
+   answer token (phase 166); the document viewer's page strip is CLASS-scoped —
   the primary class shows the strip in place of the text, the
   additive class never hides the text: Text/Pages toggle group
   (`aria-pressed`, Text default) with the strip behind it
@@ -92,6 +92,15 @@ question caps at 90 displayed characters with a `…` (phase 157) — the
 full text stays in the hover tooltip, and clicking the chip submits
 the full question.

+**No empty answer box mid-turn** (phase 166): the bubble that will hold the
+answer is created on the turn's first `thinking`/`tool` frame — it is the
+DOM anchor the Reasoning container inserts itself above — but it stays
+hidden until the **first answer token** lands in it, so an in-flight turn
+shows the typing indicator and the Reasoning bubble only, never a padded
+empty rectangle (nothing replaces the box: the existing loading feedback
+is the turn's visible state). The same rule covers a restored or shared
+answer whose persisted text is empty, which paints no box either.
+
 The document's meta row (the chat modal **and** the standalone viewer at
 `/document.html`) carries a **Copy path** button: one click copies the
 document's `source/path` (the exact form the `read` tool, the chips, and
@@ -486,14 +495,15 @@ whole git-source refresh in one click:
 ## Thinking

 The self-hosted `turbo` model reasons before it answers. That reasoning is
-streamed with the turn as `thinking` SSE events and shown in **
-collapsible "Thinking" blocks** — one block per
-contiguous thinking run, in arrival order (phase 158) — all wrapped in
-**one collapsible "Reasoning" bubble** above the answer bubble
-(phase 161): the turn renders as a chronological timeline
+streamed with the turn as `thinking` SSE events and shown as **"Thinking"
+blocks** — one block per contiguous thinking run, in arrival order
+(phase 158) — all wrapped in **one collapsible "Reasoning" bubble** above
+the answer bubble (phase 161): the turn renders as a chronological timeline
 (Thinking → Tool call (1) → Thinking → … → answer), condensed into a
 single bounded live window (320 px, one scroll surface) that follows
-the tail while you watch.
+the tail while you watch. The Reasoning bubble is the ONE disclosure
+(owner refinement, 2026-10-06): the per-run blocks inside it are plain and
+never fold on their own.

 The bubble is created on the turn's first thinking (or tool) frame and
 stays **open** from creation until the turn settles — including a
@@ -1174,11 +1184,21 @@ uv run pyright
 # Playwright E2E — install the browser once:
 uv run playwright install chromium

-# Each phase's E2E runs IN ISOLATION:
+# Each phase's E2E runs IN ISOLATION — one file per run:
 uv run pytest tests/e2e/test_import_documents.py -v --no-cov
 uv run pytest tests/e2e/test_chat_rag.py -v --no-cov
 # ...one file per phase in tests/e2e/ (the stories live in
-# .agents/user_stories/, see .agents/phases/todo/)
+# .agents/user_stories/, see .agents/phases/todo/).
+#
+# Isolation is REQUIRED, not a style preference: every suite shares one
+# Postgres and one pair of fixed ports (the app on 8123, the mock aipi on
+# 8901 — overridable with E2E_APP_PORT / E2E_MOCK_PORT), and each file
+# truncates + reseeds the KB for its own fixture. Running the whole
+# `tests/e2e/` directory in a single pytest session interleaves those
+# fixtures and fails tests that pass file by file.
+
+# To sweep every suite, loop over the files (one pytest process each):
+for f in tests/e2e/test_*.py; do uv run pytest "$f" --no-cov -q || echo "FAIL $f"; done
 ```

 **Retrieval regression battery.** `tests/fixtures/retrieval_battery.txt` is
@@ -507,6 +507,21 @@ body::before {
  overflow-wrap: anywhere;
 }
 .bubble p { margin: 0.2rem 0; }
+/* Phase 166 (TODO L4): the answer bubble an in-flight turn creates is
+   EMPTY by construction — app.js:4063/:4164 call addMessage("brain","")
+   so ensureReasoningWrap has an anchor to insert the per-run timeline
+   BEFORE. With .bubble's own padding + border + background + shadow an
+   empty one paints a ~40px box, which is the "empty box before the LLM
+   responds" the owner reported (image.png). :empty matches only a
+   child-less element — exactly innerHTML === "" (renderMarkdown("") is
+   "", so the restore and the shared page take the same path) — and the
+   first delta writes content, so the box appears AT the first token.
+   display:none also drops the flex item, so .msg-body's gap collapses.
+   The typing indicator is never :empty (three dot <span>s) — its
+   animation is untouched. Scoped to .msg .bubble: the viewer's .doc-*
+   blocks and the composer never match. Static suppression — no new
+   tokens, no animation (nothing for prefers-reduced-motion to still). */
+.msg .bubble:empty { display: none; }
 .bubble pre {
  background: var(--surface);
  color: var(--ink);
@@ -590,6 +590,37 @@ Implements just enough of the aipi surface:
    phrases — the phase-71/72/94 ordering convention); verified
    2026-09-16: no existing E2E question or fixture file contains the
    phrase, so every other suite is unaffected.
+  - user message containing ``paint the pre-answer window``
+    (``EMPTY_BUBBLE_TRIGGER``, phase 166 task 03 — the no-empty-
+    answer-box story's dedicated suite
+    ``tests/e2e/test_no_empty_box.py``) **and** the system prompt
+    carries the ``<tools>`` section -> the deterministic PRE-ANSWER
+    WINDOW turn (TODO.md L4: "There's an empty box where the LLM
+    response will go that shows up before the LLM responds"):
+    thinking frames and a tool call with ZERO content deltas — the
+    exact shape that used to paint the empty answer bubble — with
+    baked-in delays (every visible window ≥1 s — the ``slow_llm.py``
+    / phase-109 pacing precedent) so the story E2E's window
+    assertions are race-free. Discriminated statelessly from the
+    messages (streaming only):
+      * request 1 (``tools`` offered, no tool results yet): ~1.2 s
+        pre-delay (model latency while the typing indicator shows),
+        then 3 ``reasoning_content`` chunks at ~0.4 s (≈ 1.2 s of
+        thinking — this request NEVER emits ``delta.content``), then
+        the ``ls`` tool call (synthetic id ``call_0``, no arguments
+        — the phase-37 pattern), ``finish_reason: "tool_calls"``;
+      * request 2 (a ``tool``-role result in the messages — the
+        server ran the ``ls``): a ~2.0 s frameless gap (the
+        post-tool pre-answer window), then 3 short ``delta.content``
+        chunks 0.05 s apart (the answer box appears WITH the first
+        chunk), ``finish_reason: "stop"``.
+    The turn's SSE is exactly ``thinking → tool → delta → done`` —
+    the mid-turn state the box must survive hidden in (typing
+    indicator → reasoning container → answer at the first token).
+    Checked BEFORE the plain ``TOOLS_TRIGGER`` flow (disjoint trigger
+    phrases — the phase-71/72/94 ordering convention); verified
+    2026-10-08: no existing E2E question or fixture file contains
+    the phrase, so every other suite is unaffected.
  - user message containing ``repeat your folder map``
    (``FOLDER_MAP_TRIGGER``, phase 119, D3 — the suggested-folder
    context's story suite ``tests/e2e/test_name_signal_read_chips.py``)
@@ -3783,6 +3814,160 @@ def _turn_progress_stream(step: str) -> Any:
    yield "data: [DONE]\n\n"


+# ---------------------------------------------------------------------------
+# Phase 166 (task 03, the no-empty-answer-box suite): the deterministic
+# PRE-ANSWER WINDOW turn (thinking + a tool call, zero content deltas
+# until the follow-up request) — see the module docstring
+# ---------------------------------------------------------------------------
+
+#: A user message containing this substring (case-insensitive) —
+#: combined with the ``<tools>`` section in the system prompt — drives
+#: the phase-166 PRE-ANSWER WINDOW turn (TODO.md L4: "There's an
+#: empty box where the LLM response will go that shows up before the
+#: LLM responds"): ~2.4 s of thinking + a tool call with ZERO content
+#: deltas, then a ~2 s gap, then the answer — the exact shape that
+#: painted the empty answer bubble (the owner's image.png) before
+#: phase 166 hid it. Baked-in delays keep every visible window ≥1 s
+#: (the ``slow_llm.py`` / phase-109 pacing precedent) so the story
+#: E2E's window assertions are race-free. Verified 2026-10-08: no
+#: existing E2E question or fixture file contains the phrase, so
+#: every other suite is unaffected.
+EMPTY_BUBBLE_TRIGGER = "paint the pre-answer window"
+
+#: Request 1's pre-delay (s) — model latency: the typing indicator's
+#: window (no frame of the request has arrived yet).
+EMPTY_BUBBLE_PREDELAY_S = 1.2
+
+#: Per-chunk delay (s) of request 1's three ``reasoning_content``
+#: chunks — ≈ 1.2 s of visible thinking with NO ``delta.content``
+#: (the window in which the empty bubble used to paint).
+EMPTY_BUBBLE_THINK_DELAY_S = 0.4
+
+#: Request 2's frameless gap (s) — the post-tool pre-answer window:
+#: ≥ 1 s by construction, so the E2E can assert nothing paints an
+#: answer box between the ``tool`` frame and the first token.
+EMPTY_BUBBLE_TOOL_GAP_S = 2.0
+
+#: Per-chunk delay (s) of request 2's short content deltas.
+EMPTY_BUBBLE_CONTENT_DELAY_S = 0.05
+
+#: Request 1's scratchpad (byte-stable): 3 chunks —
+#: ``marker-thought-166`` is the sentinel the story E2E matches in
+#: the open reasoning container during the pre-delta window.
+EMPTY_BUBBLE_THINKING_CHUNKS = (
+    "The owner wants the exact JSON shape — ",
+    "listing the deployments folder first, ",
+    "then answering from the file itself. marker-thought-166.",
+)
+
+#: Request 2's answer (byte-stable): 3 short chunks —
+#: ``marker-emptybox-1`` is the sentinel the story E2E matches on the
+#: bubble to prove the box appeared WITH the first token.
+EMPTY_BUBBLE_ANSWER_CHUNKS = (
+    "Here is the shape from the listing: ",
+    "version 3, two record sets, zone ",
+    "Z0RESEELINK01. marker-emptybox-1.",
+)
+
+#: The full answer (the three chunks joined) — what the bubble holds
+#: once the turn settles.
+EMPTY_BUBBLE_ANSWER = "".join(EMPTY_BUBBLE_ANSWER_CHUNKS)
+
+
+def _empty_bubble_flow(body: dict[str, Any]) -> str | None:
+    """Classify a phase-166 pre-answer-window request (see the module
+    docstring). Stateless over the messages, like the other marker
+    flows:
+
+    * ``"first"``  — ``tools`` are offered and no ``tool``-role result
+      is in the messages yet: request 1 — thinking + the ``ls`` tool
+      call, ZERO content deltas.
+    * ``"second"`` — a ``tool``-role result is in the messages (the
+      server ran the ``ls``): request 2 — the ~2 s gap, then the
+      three content chunks.
+    * ``None``     — not the flow: the trigger is absent, the
+      ``<tools>`` section is missing (deflected turns never carry
+      it), or tools are not offered and no tool result is in the
+      messages yet.
+    """
+    if EMPTY_BUBBLE_TRIGGER not in _user(body).lower():
+        return None
+    if "<tools>" not in _system(body):
+        return None
+    if _tool_results(body):
+        return "second"
+    if not body.get("tools"):
+        return None
+    return "first"
+
+
+def _empty_bubble_stream(step: str) -> Any:
+    """SSE frames for one phase-166 pre-answer-window request.
+
+    ``"first"`` (model call 1): the ~1.2 s pre-delay (model latency
+    — the typing-indicator window), the 3 ``reasoning_content``
+    chunks at 0.4 s each (≈ 1.2 s of thinking — this request emits
+    NO ``delta.content`` at all: the pre-delta state that used to
+    paint the empty box), the ``ls`` tool-call partial (synthetic id
+    ``call_0``, no arguments — the phase-37 ``_tool_call_stream``
+    first-partial shape), and the ``finish_reason: "tool_calls"``
+    frame.
+
+    ``"second"`` (model call 2 — after the server ran the ``ls``):
+    the ~2.0 s frameless gap (the post-tool pre-answer window), the
+    3 short ``delta.content`` chunks 0.05 s apart (the box appears
+    with the first chunk and grows from there), and the
+    ``finish_reason: "stop"`` frame.
+    """
+    model = "turbo"
+    chunk_id = f"chatcmpl-{uuid.uuid4()}"
+
+    def frame(delta: dict[str, Any], finish: str | None = None) -> str:
+        return (
+            "data: "
+            + json_dumps(
+                {
+                    "id": chunk_id,
+                    "object": "chat.completion.chunk",
+                    "created": int(time.time()),
+                    "model": model,
+                    "choices": [{"index": 0, "delta": delta, "finish_reason": finish}],
+                }
+            )
+            + "\n\n"
+        )
+
+    if step == "first":
+        time.sleep(EMPTY_BUBBLE_PREDELAY_S)
+        for piece in EMPTY_BUBBLE_THINKING_CHUNKS:
+            yield frame({"reasoning_content": piece})
+            time.sleep(EMPTY_BUBBLE_THINK_DELAY_S)
+        yield frame(
+            {
+                "role": "assistant",
+                "tool_calls": [
+                    {
+                        "index": 0,
+                        "id": "call_0",
+                        "type": "function",
+                        "function": {"name": "ls", "arguments": "{}"},
+                    }
+                ],
+            }
+        )
+        time.sleep(0.1)
+        yield frame({}, "tool_calls")
+        yield "data: [DONE]\n\n"
+        return
+
+    time.sleep(EMPTY_BUBBLE_TOOL_GAP_S)
+    for piece in EMPTY_BUBBLE_ANSWER_CHUNKS:
+        yield frame({"content": piece})
+        time.sleep(EMPTY_BUBBLE_CONTENT_DELAY_S)
+    yield frame({}, "stop")
+    yield "data: [DONE]\n\n"
+
+
 def compose_answer(body: dict[str, Any]) -> str:
    system = _system(body)
    user = _user(body)
@@ -5053,6 +5238,23 @@ def chat_completions(body: dict[str, Any]) -> Any:
                media_type="text/event-stream",
                headers={"Cache-Control": "no-cache", "X-Accel-Buffering": "no"},
            )
+        # Phase 166 (task 03, the no-empty-answer-box suite): the
+        # deterministic PRE-ANSWER WINDOW turn — request 1 streams
+        # thinking + the ``ls`` tool call with ZERO content deltas,
+        # request 2 answers after a ~2 s gap (TODO.md L4 — the empty
+        # answer box that painted before the first token; see the
+        # module docstring). Checked BEFORE the plain TOOLS_TRIGGER
+        # flow (disjoint trigger phrases — the phase-71/72/94
+        # ordering convention; the trigger needs the ``<tools>``
+        # section, so deflected turns never hit it).
+        empty_bubble = _empty_bubble_flow(body)
+        if empty_bubble is not None:
+            stream = _empty_bubble_stream(empty_bubble)
+            return StreamingResponse(
+                stream,
+                media_type="text/event-stream",
+                headers={"Cache-Control": "no-cache", "X-Accel-Buffering": "no"},
+            )
        # Phase 161 (task 04 — the reasoning-bubble story): the
        # deterministic MID-ROUND flow — a thinking run, then a
        # MID-ROUND ``delta`` (the React-style partial answer — the
@@ -521,7 +521,7 @@ def _expect_desc_cell(
    escape hatch, ``makeDescCell`` — set whenever the description is
    non-empty, whether or not the display ellipsizes)."""
    span = page.locator(f"#folders-tbody tr:nth-child({row_index + 1}) "
-                       "td:nth-child(4) span")
+                       "td:nth-child(4) span.kb-desc-text")
    expect(span).to_have_text(full_text)
    expect(span).to_have_attribute("title", full_text)

@@ -633,7 +633,7 @@ def test_rag_table_renders_the_stored_voice(
    expect(page.locator("#kb-level-summary")).to_have_text(EXPECTED[""])
    # The rendered voice is the new one: no Description cell in the
    # level shows a forbidden opener (the shape pin, view-side).
-    for cell in page.locator("#folders-tbody td:nth-child(4) span").all():
+    for cell in page.locator("#folders-tbody td:nth-child(4) span.kb-desc-text").all():
        rendered = cell.text_content() or ""
        for prefix in FORBIDDEN_PREFIXES:
            assert not rendered.startswith(prefix), rendered
@@ -91,6 +91,33 @@ def _box_height(page: Page, selector: str) -> float:
    return box["height"]


+def _meta_line_count(page: Page, selector: str) -> int:
+    """How many VISUAL LINES the row's children occupy: cluster the
+    children by their vertical center (a 4px tolerance absorbs the
+    mixed badge/text/native-control heights on one line) and return the
+    number of clusters. ONE cluster = one line; a wrap produces two (the
+    second row of children sits a full line lower). Shape-first on
+    purpose: the row's own height is content-sized by its tallest child
+    — since phase 153 that is the ≥44px touch-target floor on
+    `.doc-copy-path` at ≤640px, so a pixel bound on the row would pin
+    the control's minimum, not the layout."""
+    centers = page.evaluate(
+        """(sel) => [...document.querySelectorAll(sel + ' > *')]
+            .map(el => { const b = el.getBoundingClientRect();
+                         return b.height > 0 ? b.top + b.height / 2 : null; })
+            .filter(v => v !== null)""",
+        selector,
+    )
+    assert centers, f"{selector} rendered no visible children"
+    clusters: list[list[float]] = []
+    for c in sorted(centers):
+        if clusters and c - clusters[-1][-1] <= 4:
+            clusters[-1].append(c)
+        else:
+            clusters.append([c])
+    return len(clusters)
+
+
 def _header_heights(page: Page, app_url: str) -> dict[str, float]:
    """Measured heights of the sticky top bar on the three pages."""
    heights: dict[str, float] = {}
@@ -201,13 +228,19 @@ def test_viewer_header_content_still_fits(
        assert _box_height(page, "#doc-title") < 30  # one line at either size

        # Meta row: badges + path present and visible, single line.
-        # The bound catches a WRAP (a second line would be ≥ ~44px),
-        # not a pixel-exact line height: the row is content-sized by
-        # its tallest child — the native type=date input (phase 106
-        # D8), whose metrics are Chromium/font-dependent and render
-        # the single line at ~26px on this host (the pre-phase 26px
-        # pin red-lined on the 26.125px measurement, 2026-09-16).
-        assert _box_height(page, ".doc-meta") < 34
+        # SHAPE, not pixels: the pre-phase absolute bound (<34px) went
+        # stale the moment the row gained a control carrying the ≥44px
+        # touch-target floor (phase 153: .doc-copy-path min-height 44px
+        # at ≤640px — PLAN §7.1), which makes a single, correctly-wrapped
+        # mobile line measure 44px. What the contract actually forbids is
+        # the row STACKING (a wrap grows the header), so the pin reads the
+        # geometry directly: every child centered on the row's own center
+        # = ONE line, whatever the tallest control's height (the same
+        # shape-first reading the 2026-09-16 note above already moved to).
+        assert _meta_line_count(page, ".doc-meta") == 1, (
+            "the meta row must stay on ONE line (badges, path, dates and "
+            "controls centered on the row's own center)"
+        )
        expect(page.locator(".doc-source-badge", has_text="docs")).to_be_visible()
        expect(page.locator(".format-badge", has_text="md")).to_be_visible()
        expect(page.locator(".doc-path", has_text="homelab/kubernetes.md")).to_be_visible()
@@ -545,10 +545,10 @@ def test_top_level_lists_sources_with_descriptions(
    rows = page.locator("#folders-tbody tr")
    expect(rows.nth(0).locator("a.folder-link")).to_have_text(ALPHA)
    expect(rows.nth(0).locator("td:nth-child(2)")).to_have_text(str(ALPHA_COUNT))
-    expect(rows.nth(0).locator("td:nth-child(4) span")).to_have_text(ALPHA_ROOT_SUM)
+    expect(rows.nth(0).locator("td:nth-child(4) span.kb-desc-text")).to_have_text(ALPHA_ROOT_SUM)
    expect(rows.nth(1).locator("a.folder-link")).to_have_text(BETA)
    expect(rows.nth(1).locator("td:nth-child(2)")).to_have_text(str(BETA_COUNT))
-    expect(rows.nth(1).locator("td:nth-child(4) span")).to_have_text(BETA_ROOT_SUM)
+    expect(rows.nth(1).locator("td:nth-child(4) span.kb-desc-text")).to_have_text(BETA_ROOT_SUM)
    # The edit affordance is ALWAYS present (a description can be
    # CREATED where none is stored) — the row surface, one per row.
    expect(page.locator("#folders-tbody .kb-summary-edit")).to_have_count(2)
@@ -607,10 +607,10 @@ def test_drill_into_source(
    expect(rows).to_have_count(2)
    expect(rows.nth(0).locator("a.folder-link")).to_have_text("one")
    expect(rows.nth(0).locator("td:nth-child(2)")).to_have_text("2")
-    expect(rows.nth(0).locator("td:nth-child(4) span")).to_have_text(ONE_SUM)
+    expect(rows.nth(0).locator("td:nth-child(4) span.kb-desc-text")).to_have_text(ONE_SUM)
    expect(rows.nth(1).locator("a.folder-link")).to_have_text("two")
    expect(rows.nth(1).locator("td:nth-child(2)")).to_have_text("2")
-    expect(rows.nth(1).locator("td:nth-child(4) span")).to_have_text(TWO_SUM)
+    expect(rows.nth(1).locator("td:nth-child(4) span.kb-desc-text")).to_have_text(TWO_SUM)

    # The level's direct files: root-note.md — the 6-column contract
    # (makeRow): Source | Path | Title | Chunks | Created | Indexed
@@ -718,7 +718,7 @@ def test_edit_folder_description(
    # (the textContent re-render — the canned text is gone); the
    # always-present Edit button is back.
    expect(status).to_have_text("Description updated.")
-    expect(row.locator("td:nth-child(4) span")).to_have_text(NEW_ALPHA_ONE)
+    expect(row.locator("td:nth-child(4) span.kb-desc-text")).to_have_text(NEW_ALPHA_ONE)
    expect(row.locator(".kb-summary-edit")).to_be_visible()

    # The server row: the new text AND the manually_edited flag (the
@@ -761,7 +761,7 @@ def test_clear_folder_description(
    # The text is gone: the cell empties (the always-present Edit
    # button stays — a description can be re-created from the empty
    # cell).
-    expect(row.locator("td:nth-child(4) span")).to_have_text("")
+    expect(row.locator("td:nth-child(4) span.kb-desc-text")).to_have_text("")
    expect(row.locator(".kb-summary-edit")).to_be_visible()

    # The SQL assert: no row for the folder (the next KB-changing
@@ -794,7 +794,7 @@ def test_manual_description_survives_a_changed_sync(
    rows = page.locator("#folders-tbody tr")
    expect(rows).to_have_count(1)
    expect(rows.nth(0).locator("a.folder-link")).to_have_text("gamma")
-    expect(rows.nth(0).locator("td:nth-child(4) span")).to_have_text(GAMMA_SUM)
+    expect(rows.nth(0).locator("td:nth-child(4) span.kb-desc-text")).to_have_text(GAMMA_SUM)

    # The LEVEL-BLOCK Edit (the level surface — test 4 pins the row
    # surface): the static #kb-level-edit button, the source root
@@ -836,7 +836,7 @@ def test_manual_description_survives_a_changed_sync(
    expect(
        page.locator("#folders-tbody tr")
        .nth(0)
-        .locator("td:nth-child(4) span")
+        .locator("td:nth-child(4) span.kb-desc-text")
    ).to_have_text(GAMMA_SUM)

    # The new document landed in the tree: back to the top (the
@@ -847,7 +847,7 @@ def test_manual_description_survives_a_changed_sync(
    one_row = page.locator("#folders-tbody tr").nth(0)
    expect(one_row.locator("a.folder-link")).to_have_text("one")
    expect(one_row.locator("td:nth-child(2)")).to_have_text("3")
-    expect(one_row.locator("td:nth-child(4) span")).to_have_text(ONE_SUM)
+    expect(one_row.locator("td:nth-child(4) span.kb-desc-text")).to_have_text(ONE_SUM)

    # The SQL asserts: the manual row survived (text + flag), the
    # untouched rows were regenerated (the canned text, the flag back
@@ -0,0 +1,713 @@
+"""Phase 166 E2E (Playwright, mock-only): NO empty answer box paints
+before the first token (owner TODO.md L4 — "There's an empty box where
+the LLM response will go that shows up before the LLM responds. It
+looks ugly." — the attached image.png: Thinking ▸ / Tool calls (2) ▸ /
+an expanded Thinking run, then a bordered, padded, EMPTY rounded box).
+
+Run in isolation (DB must be up: ``podman compose up -d db``):
+
+    uv run pytest tests/e2e/test_no_empty_box.py -v --no-cov
+
+MOCK-ONLY suite: the gate is the deterministic paced marker flow in
+``tests/e2e/mock_llm.py`` (user message contains ``paint the pre-answer
+window`` (``EMPTY_BUBBLE_TRIGGER``) **and** the system prompt carries
+the ``<tools>`` section of the HIGH prompt): request 1 streams ~1.2 s
+of ``reasoning_content`` and one ``ls`` tool call with ZERO content
+deltas (the exact mid-turn state of the owner's screenshot), request 2
+(id a ``tool``-result) waits ~2.0 s, then answers in three short
+chunks. Every visible window is ≥ 1 s by construction (the phase-109
+pacing precedent), so the window assertions are race-free — the
+pre-delta window from the first thinking frame to the first delta is
+~3.2 s.
+
+The fix is the ONE CSS rule ``.msg .bubble:empty { display: none; }``
+(LOCKED A1 — hide in CSS, the JS keeps creating the empty bubble as
+the phase-158/161 reasoning-container anchor); nothing replaces the
+box (A2 — the typing indicator + the reasoning container ARE the
+mid-turn feedback, PLAN §7.4), and the same rule covers the restore
+and the shared page (A3). The unit pins live in
+``tests/unit/test_no_empty_answer_bubble.py``; this file proves the
+browser behavior.
+
+KB fixture — the deterministic two-document pair mirrored from
+``test_reasoning_bubble.py``: the marker question cosines well past
+the mock-calibrated 0.30 E2E threshold plus FTS hits, so the turn is
+solidly grounded (HIGH prompt → ``<tools>``).
+
+Test → phase mapping (Playwright Mapping Rule):
+1. ``test_no_box_paints_before_the_first_token`` — the defect window
+   (D1/D3): once ``details.reasoning`` exists (the sync point — the
+   first thinking frame, which also removes the typing indicator),
+   ``.msg.brain .bubble`` is NOT visible and its box is zero-height
+   while the thinking + tool frames stream, while the reasoning
+   container IS open with a live status literal (``Thinking…`` /
+   ``Calling ``) and the composer reads the in-flight state (``Stop``).
+   The mid-turn screenshot is the owner's before/after record. THE
+   SETTLE: after ``done`` the SAME locator is visible with ``height >
+   0`` holding the answer, the container folds, and the phase-130 rows
+   (chips, actions) sit BELOW the bubble (hiding the bubble never
+   hides them).
+2. ``test_plain_turn_typing_indicator_carries_the_window`` — the
+   no-thinking control: a plain answer question shows the typing
+   indicator (dots ⇒ never ``:empty`` ⇒ painted with height > 0)
+   while in flight, and NO empty bubble ever paints a box. The plain
+   turn is FAST (the mock streams unpaced), so the proof is an
+   IN-PAGE sampler on a 10 ms interval installed BEFORE the click:
+   typing painted ≥ 1 sample, painted empty boxes 0 samples across
+   every sample of the turn (the answer bubble is created WITH its
+   first delta in one JS task — nothing can paint between them, and
+   the sampler is the browser-side evidence of exactly that; A2).
+3. ``test_restored_answered_chat_has_no_empty_box`` — the restore
+   corner (A3): the auto-saved chat re-loaded through ``/?chat=<id>``
+   renders its answer bubble normally and carries ZERO
+   ``.bubble:empty`` elements.
+4. ``test_shared_page_has_no_empty_box`` — the shared-page corner
+   (A3): the phase-51 save + share flow, a fresh anonymous context on
+   ``/shared/<token>`` — the answer renders, zero ``.bubble:empty``.
+5. ``test_empty_persisted_answer_paints_no_box`` — the empty-TEXT
+   record corner (A3, the shared-page defect half): a saved chat
+   whose persisted brain text is whitespace (passes the client's
+   ``text.length > 0`` filter, renders through ``renderMarkdown(" ")
+   === ""``) paints NO box on ``/?chat=<id>`` AND on its shared
+   page — the bubble exists in the DOM (the anchor contract) but
+   zero-height and hidden.
+"""
+from __future__ import annotations
+
+import hashlib
+import re
+import time
+import uuid
+from collections.abc import Callable
+from datetime import UTC, datetime
+from pathlib import Path
+from typing import Any
+
+import httpx
+from playwright.sync_api import Browser, BrowserContext, Page, expect
+from sqlalchemy import text
+from sqlalchemy.orm import Session
+
+from app.db import SessionLocal
+from app.models import Chunk, Document, GitSource, SavedChat
+from e2e.auth_helpers import login
+from tests.e2e.mock_llm import EMPTY_BUBBLE_ANSWER, EMPTY_BUBBLE_TRIGGER, embed_text
+
+REPO = Path(__file__).resolve().parents[2]
+SCREENSHOTS = REPO / ".agents" / "screenshots"  # house convention for visual records
+
+# --------------------------------------------------------------------------
+# Fixture documents (deterministic, token-controlled — mirrored from
+# test_reasoning_bubble.py so the grounding gate is byte-stable)
+# --------------------------------------------------------------------------
+
+SEED_SOURCE = "Homelab"
+SEED_PATH = "aws-route53.md"
+
+READ_SOURCE = "Deployments"
+READ_PATH = "example-record-file.json"
+
+ROUTE53_CONTENT = (
+    "# AWS Route 53 Notes\n\n"
+    "## Record file\n\n"
+    + (
+        "The aws route53 hosted zone for reeselink keeps every record in "
+        "reseelink.json — the exact JSON shape of reeselink.json is "
+        "documented in example-record-file.json.\n"
+    )
+    * 10
+    + "\n## Sync job\n\n"
+    "A cron job pushes reeselink.json to the aws route53 hosted zone "
+    "every fifteen minutes; the diff is applied through the route53 api.\n"
+)
+
+RECORD_FILE_CONTENT = (
+    '{ "version": 3, "comment": "ReeseLink hosted zone records — the exact '
+    "JSON shape of reeselink.json\",\n"
+    '  "hosted_zone_id": "Z0RESEELINK01",\n'
+    '  "record_sets": [\n'
+    '    { "name": "www.reeselink.example", "type": "A", "ttl": 300,\n'
+    '      "resource_records": [ { "value": "10.0.0.20" } ] },\n'
+    '    { "name": "api.reeselink.example", "type": "CNAME", "ttl": 300,\n'
+    '      "resource_records": [ { "value": "www.reeselink.example" } ] }\n'
+    "  ]\n"
+    "}\n"
+)
+
+#: The phase-166 marker question: ``EMPTY_BUBBLE_TRIGGER`` + the KB
+#: topic — the mock-calibrated cosine keeps the turn grounded (HIGH
+#: prompt → ``<tools>`` → the deterministic pre-answer-window flow).
+EMPTY_WINDOW_QUESTION = (
+    f"{EMPTY_BUBBLE_TRIGGER.capitalize()}: what is the exact JSON shape "
+    "of reeselink.json for my aws route53 hosted zone?"
+)
+
+#: The no-thinking control (NO trigger — the generic ``compose_answer``
+#: path: no thinking frames, no tool frames; the typing indicator is
+#: the whole pre-answer state).
+PLAIN_ANSWER_QUESTION = (
+    "What is the exact JSON shape of reeselink.json for my aws "
+    "route53 hosted zone?"
+)
+
+#: The live status literals of the reasoning container (phase 161 D6)
+#: — the thinking/tool state's visible text while the box stays hidden.
+LIVE_STATUS_RE = re.compile(r"^(Thinking…|Calling )")
+
+SHARE_URL_RE = re.compile(r"^/shared/[0-9a-f-]{36}$")
+
+# --------------------------------------------------------------------------
+# DB seeding (TRUNCATE-then-seed, mirrored from test_reasoning_bubble.py)
+# --------------------------------------------------------------------------
+
+
+def _seed(db: Session) -> None:
+    """The two-document pair (see the module docstring), plus the
+    phase-94 registry rows: both sources registered, ``Deployments``
+    FIRST (the operator's ``BOR_GIT_SOURCES`` is ignored — a non-empty
+    table wins). COMMIT between the inserts (not flush): ``added_at``
+    is ``server_default now()`` and the tie-break is the random uuid
+    ``id`` — one-transaction rows order nondeterministically."""
+    db.add(GitSource(url=READ_SOURCE, kind="local"))
+    db.commit()
+    db.add(GitSource(url=SEED_SOURCE, kind="local"))
+    md = Document(
+        source=SEED_SOURCE,
+        path=SEED_PATH,
+        full_path=f"/tmp/{SEED_PATH}",
+        title="AWS Route 53 Notes",
+        content=ROUTE53_CONTENT,
+        content_hash=hashlib.sha256(ROUTE53_CONTENT.encode()).hexdigest(),
+        indexed_at=datetime.now(UTC),
+        created_at=datetime(2024, 6, 15, tzinfo=UTC),
+    )
+    db.add(md)
+    db.flush()
+    # One chunk carrying the mock's own embedding → genuine token
+    # overlap between the marker question and this document.
+    db.add(
+        Chunk(
+            document_id=md.id,
+            position=0,
+            content=ROUTE53_CONTENT,
+            embedding=embed_text(ROUTE53_CONTENT),
+        )
+    )
+    db.add(
+        Document(
+            source=READ_SOURCE,
+            path=READ_PATH,
+            full_path=f"/tmp/{READ_PATH}",
+            title="Example Record File",
+            content=RECORD_FILE_CONTENT,
+            content_hash=hashlib.sha256(RECORD_FILE_CONTENT.encode()).hexdigest(),
+            indexed_at=datetime.now(UTC),
+            created_at=datetime(2024, 6, 15, tzinfo=UTC),
+        )
+    )
+
+
+def _reset_db(seed: Callable[[Session], None] | None = None) -> None:
+    """Truncate the KB (plus the prompt-shaping tables and the
+    phase-55 auto-save rows), then re-seed — the E2E isolation
+    pattern. ``steering_notes`` / ``kb_overview`` are truncated too,
+    so the HIGH prompt is exactly ``<relevance>`` + ``<documents>`` +
+    ``<tools>`` regardless of leftovers — byte-stable prompts,
+    byte-stable answers."""
+    with SessionLocal() as db:
+        db.execute(
+            text(
+                "TRUNCATE chunks, documents, query_log, "
+                "steering_notes, kb_overview, saved_chats, git_sources"
+            )
+        )
+        db.commit()
+        if seed is not None:
+            seed(db)
+            db.commit()
+
+
+# --------------------------------------------------------------------------
+# Page helpers
+# --------------------------------------------------------------------------
+
+#: The defect's visible signature: a ``.msg .bubble`` element that is
+#: EMPTY (child-less — exactly what ``addMessage("brain","")``
+#: produces) yet still PAINTS A BOX (height > 0). The owner's
+#: image.png is this number being ≥ 1; phase 166 makes it 0 at every
+#: sample, live, restored, and shared (the same stylesheet on the
+#: shared page — LOCKED A3).
+PAINTED_EMPTY_BOXES_JS = """
+() => [...document.querySelectorAll('#messages .msg .bubble')]
+    .filter((el) => el.matches(':empty') && el.getBoundingClientRect().height > 0)
+    .length
+"""
+
+#: The IN-PAGE sampler (the no-thinking control's race-free proof):
+#: a 10 ms interval installed BEFORE the submit counts (a) samples
+#: where ``#typing-indicator .bubble`` painted a box and (b) samples
+#: where any EMPTY ``.msg .bubble`` painted a box. The plain turn can
+#: settle in ~100 ms — too fast for out-of-band polling to bracket
+#: reliably; the sampler sees every 10 ms slice of the page from the
+#: click to the settle (and a few past it) from INSIDE the page.
+SAMPLER_ON_JS = """
+() => {
+  window.__p166 = { samples: 0, typingPainted: 0, paintedEmpty: 0 };
+  window.__p166.timer = setInterval(() => {
+    const s = window.__p166;
+    s.samples += 1;
+    const t = document.querySelector('#typing-indicator .bubble');
+    if (t && t.getBoundingClientRect().height > 0) s.typingPainted += 1;
+    const empties = [...document.querySelectorAll('#messages .msg .bubble')]
+      .filter((el) => el.matches(':empty')
+        && el.getBoundingClientRect().height > 0);
+    if (empties.length > 0) s.paintedEmpty += 1;
+  }, 10);
+}
+"""
+
+SAMPLER_OFF_JS = """
+() => {
+  clearInterval(window.__p166.timer);
+  const { samples, typingPainted, paintedEmpty } = window.__p166;
+  return { samples, typingPainted, paintedEmpty };
+}
+"""
+
+
+def _bubble_height(locator: Any) -> float:
+    """The painted box height of one bubble — ``display: none`` (the
+    phase-166 rule) yields exactly 0."""
+    height: float = locator.evaluate("el => el.getBoundingClientRect().height")
+    return height
+
+
+def _submit(page: Page, question: str) -> None:
+    page.fill("#message-input", question)
+    page.click("#send-btn")
+    # The user bubble lands synchronously with the submit handler.
+    expect(page.locator(".msg.user .bubble").last).to_contain_text(question)
+
+
+def _wait_settled(page: Page) -> None:
+    """The turn is complete: answer text in the bubble, button
+    recovered (phase 48 — the in-flight button is the enabled Stop
+    control, so the label assertion carries the settle wait)."""
+    expect(page.locator(".msg.brain .bubble").last).not_to_have_text("", timeout=30_000)
+    expect(page.locator("#send-btn")).to_be_enabled(timeout=30_000)
+    expect(page.locator("#send-label")).to_have_text("Send", timeout=30_000)
+
+
+# --------------------------------------------------------------------------
+# Phase-51 save + share flow (the test_reasoning_bubble.py pattern — the
+# auto-saved row is shared from the chat page; the row is deleted in a
+# finally)
+# --------------------------------------------------------------------------
+
+
+def _admin_cookies(page: Page) -> dict[str, str]:
+    return {
+        c["name"]: c["value"]
+        for c in page.context.cookies()
+        if "name" in c and "value" in c
+    }
+
+
+def _chats(app_url: str, cookies: dict[str, str]) -> list[dict[str, Any]]:
+    r = httpx.get(f"{app_url}/api/chats", timeout=10, cookies=cookies)
+    assert r.status_code == 200
+    return r.json()["chats"]
+
+
+def _auto_title(question: str) -> str:
+    """The phase-50 auto-title convention: the first question,
+    whitespace-collapsed, capped at 120 chars."""
+    return " ".join(question.split())[:120]
+
+
+def _find_row(rows: list[dict[str, Any]], title: str) -> dict[str, Any] | None:
+    return next((c for c in rows if c["title"] == title), None)
+
+
+def _delete_chat(app_url: str, cookies: dict[str, str], chat_id: str) -> None:
+    """Best-effort row cleanup (a 404 — already deleted — is fine)."""
+    httpx.delete(f"{app_url}/api/chats/{chat_id}", timeout=10, cookies=cookies)
+
+
+def _wait_saved_row(
+    app_url: str,
+    cookies: dict[str, str],
+    title: str,
+    messages: int = 2,
+) -> dict[str, Any]:
+    """Wait for the auto-saved row (phase 55 — auto-saves are SILENT,
+    so there is no status line to wait on). The upsert is
+    fire-and-forget from the UI's point of view, so poll the admin
+    list until the row with the conversation's auto-title appears
+    with the expected message count."""
+    deadline = time.monotonic() + 15
+    last: dict[str, Any] | None = None
+    while time.monotonic() < deadline:
+        last = _find_row(_chats(app_url, cookies), title)
+        if last is not None and last["message_count"] >= messages:
+            return last
+        time.sleep(0.2)
+    raise AssertionError(f"no auto-saved row for {title!r} (last: {last!r})")
+
+
+def _grant_clipboard(page: Page, app_url: str) -> None:
+    """Grant the async-clipboard permissions on the admin context
+    (headless Chromium requires the grant before ``writeText``
+    resolves — without it the inline-link fallback fires)."""
+    page.context.grant_permissions(
+        ["clipboard-read", "clipboard-write"], origin=app_url
+    )
+
+
+def _click_share_and_assert_status(page: Page, app_url: str) -> None:
+    """Press the chat page's Share pill and pin the outcome on the
+    live region (the assertion branches on the API's availability —
+    the test_share_chat.py convention)."""
+    page.locator("#share-chat-btn").click()
+    if page.evaluate("() => !!navigator.clipboard"):
+        expect(page.locator("#send-status")).to_have_text(
+            "Share link copied.", timeout=15_000
+        )
+    else:
+        expect(page.locator("#send-status")).to_have_text(
+            "Share link ready — copy it from the field.", timeout=15_000
+        )
+        field = page.locator(".share-link-fallback")
+        expect(field).to_be_visible()
+        href = field.get_attribute("href")
+        assert href is not None
+        assert href.startswith(app_url)
+        assert SHARE_URL_RE.fullmatch(href.removeprefix(app_url))
+
+
+# --------------------------------------------------------------------------
+# 1. The defect window: thinking + tool frames stream and NO answer
+#    box paints — the box appears WITH the first delta (D1/D3)
+# --------------------------------------------------------------------------
+
+
+def test_no_box_paints_before_the_first_token(
+    page: Page, app_url: str, mock_llm: int, db_ready: None
+) -> None:
+    page.set_default_timeout(30_000)
+    _reset_db(_seed)
+    login(page, app_url, next="/")
+
+    _submit(page, EMPTY_WINDOW_QUESTION)
+
+    # THE SYNC POINT: details.reasoning exists only after the FIRST
+    # thinking frame (ensureReasoningWrap runs in the same handler
+    # tick as addMessage("brain", "") and removeTyping()) — so from
+    # here on the ONLY .msg.brain .bubble is the empty answer bubble,
+    # and the mock's pacing keeps the window open ≥ 1 s of thinking +
+    # the tool round + the ~2 s gap (≈ 3.2 s total before the first
+    # delta — race-free by construction).
+    expect(page.locator("#messages .msg.brain details.reasoning")).to_be_visible()
+
+    # THE DEFECT WINDOW: the answer bubble exists in the DOM (the
+    # phase-158/161 reasoning anchor — LOCKED A1 keeps it) but paints
+    # NO BOX: not visible AND zero-height (the owner's image.png box
+    # is gone). The typing indicator is gone too (the live block
+    # replaced the dots), so the single strict match doubles as the
+    # no-stray-bubble pin.
+    bubble = page.locator("#messages .msg.brain .bubble")
+    expect(bubble).not_to_be_visible()
+    assert _bubble_height(bubble) == 0, (
+        "the pre-delta answer bubble must paint zero box (the phase-166 defect)"
+    )
+    assert page.evaluate(PAINTED_EMPTY_BOXES_JS) == 0, (
+        "no empty .bubble may paint a box mid-turn (live window)"
+    )
+
+    # §7.4 — the mid-turn feedback DOES paint: the reasoning container
+    # is OPEN with a thinking run and its live status literal
+    # (Thinking… / Calling <name>…), and the composer carries the
+    # in-flight state (the phase-48 Stop label).
+    expect(page.locator("#messages .msg.brain details.reasoning[open]")).to_be_visible()
+    expect(
+        page.locator("#messages .msg.brain details.reasoning .thinking-text")
+    ).to_be_visible()
+    expect(
+        page.locator("#messages .msg.brain details.reasoning .reasoning-status")
+    ).to_have_text(LIVE_STATUS_RE)
+    expect(page.locator("#send-label")).to_have_text("Stop")
+
+    # The owner's before/after record (image.png's twin): the open
+    # reasoning run with NO empty box under it.
+    SCREENSHOTS.mkdir(parents=True, exist_ok=True)
+    shot = SCREENSHOTS / "166_pre_answer_window_1280.png"
+    page.screenshot(path=str(shot))
+    assert shot.exists() and shot.stat().st_size > 0
+
+    # THE SETTLE: the box appears at (and after) the first token. The
+    # SAME locator is now visible with a real box and holds the
+    # byte-stable answer (the three chunks of the mock's request 2).
+    _wait_settled(page)
+    expect(bubble).to_be_visible()
+    assert _bubble_height(bubble) > 0, "the answer box must paint with the first token"
+    expect(bubble).to_contain_text(EMPTY_BUBBLE_ANSWER)
+
+    # The container folded at done (phase 161 D2 — the settle fold).
+    expect(page.locator("#messages .msg.brain details.reasoning[open]")).to_have_count(0)
+
+    # Phase 130 ordering survives the hidden box: the chips row and
+    # the actions row are SIBLING rows of .msg-body BELOW the bubble
+    # (hiding an empty bubble never hides a chip or an action).
+    order = page.evaluate(
+        """
+        () => {
+          const wraps = document.querySelectorAll('#messages .msg.brain');
+          const wrap = wraps[wraps.length - 1];
+          const body = wrap ? wrap.querySelector('.msg-body') : null;
+          const kids = body ? [...body.children] : [];
+          return {
+            bubbleIdx: kids.findIndex((k) => k.classList.contains('bubble')),
+            metaIdxs: kids
+              .map((k, i) => (k.classList.contains('msg-meta') ? i : -1))
+              .filter((i) => i >= 0),
+          };
+        }
+        """
+    )
+    assert order["bubbleIdx"] >= 0, "the settled turn keeps its answer bubble"
+    assert order["metaIdxs"], "the settled turn keeps its .msg-meta rows"
+    assert all(idx > order["bubbleIdx"] for idx in order["metaIdxs"]), (
+        f"every .msg-meta row sits BELOW the bubble: {order}"
+    )
+    expect(
+        page.locator("#messages .msg.brain .msg-meta.msg-actions")
+    ).to_have_count(1)
+
+
+# --------------------------------------------------------------------------
+# 2. The no-thinking control: the typing indicator (dots ⇒ never
+#    :empty) carries the pre-answer window; no empty box at any point
+# --------------------------------------------------------------------------
+
+
+def test_plain_turn_typing_indicator_carries_the_window(
+    page: Page, app_url: str, mock_llm: int, db_ready: None
+) -> None:
+    page.set_default_timeout(30_000)
+    _reset_db(_seed)
+    login(page, app_url, next="/")
+
+    # The plain turn is FAST (the mock streams unpaced — the whole
+    # turn can settle in ~100 ms), so the proof runs IN-PAGE: the
+    # sampler counts 10 ms slices from BEFORE the click to a few
+    # hundred ms past the settle. The typing indicator lands
+    # synchronously in the click handler (setUiState(thinking) →
+    # addTyping(): three dot <span>s ⇒ never :empty ⇒ painted) and
+    # the answer bubble is created WITH its first delta in one JS
+    # task — every slice of the turn is accounted for (locked A2:
+    # this affordance IS the pre-answer window).
+    page.evaluate(SAMPLER_ON_JS)
+
+    _submit(page, PLAIN_ANSWER_QUESTION)
+    _wait_settled(page)
+    # A few extra samples past the settle (the painted answer bubble
+    # is non-empty — the sampler must still read zero painted boxes).
+    time.sleep(0.3)
+    stats = page.evaluate(SAMPLER_OFF_JS)
+
+    assert stats["samples"] >= 5, f"the sampler never ran: {stats}"
+    assert stats["typingPainted"] >= 1, (
+        "the typing indicator was never observed painting its dots box "
+        f"while the turn was in flight: {stats}"
+    )
+    assert stats["paintedEmpty"] == 0, (
+        f"an empty answer box painted during the plain turn: {stats}"
+    )
+
+    expect(page.locator(".msg.brain .bubble").last).not_to_have_text("")
+    assert page.evaluate(PAINTED_EMPTY_BOXES_JS) == 0
+
+
+# --------------------------------------------------------------------------
+# 3. The restore corner: an auto-saved chat re-loaded through
+#    /?chat=<id> renders its bubble normally, zero .bubble:empty
+# --------------------------------------------------------------------------
+
+
+def test_restored_answered_chat_has_no_empty_box(
+    page: Page, app_url: str, mock_llm: int, db_ready: None
+) -> None:
+    page.set_default_timeout(30_000)
+    _reset_db(_seed)
+    login(page, app_url, next="/")
+
+    _submit(page, PLAIN_ANSWER_QUESTION)
+    _wait_settled(page)
+
+    cookies = _admin_cookies(page)
+    row = _wait_saved_row(app_url, cookies, _auto_title(PLAIN_ANSWER_QUESTION))
+    try:
+        # The phase-50 restore: /?chat=<id> boots into the saved
+        # conversation through renderStoredMessage — the answered turn
+        # renders its bubble normally (non-empty, painted).
+        page.goto(app_url + f"/?chat={row['id']}")
+        expect(page.locator("#empty-state")).to_be_hidden(timeout=30_000)
+        bubble = page.locator(".msg.brain .bubble").last
+        expect(bubble).to_be_visible()
+        expect(bubble).not_to_have_text("")
+        assert _bubble_height(bubble) > 0
+
+        # ZERO child-less bubbles exist in the restored DOM — the
+        # restore creates no empty anchor (the anchor is a LIVE-turn
+        # construct), so nothing even needs hiding.
+        assert page.locator("#messages .msg .bubble:empty").count() == 0
+        assert page.evaluate(PAINTED_EMPTY_BOXES_JS) == 0
+    finally:
+        _delete_chat(app_url, cookies, row["id"])
+
+
+# --------------------------------------------------------------------------
+# 4. The shared-page corner (LOCKED A3): the phase-51 share flow, a
+#    fresh anonymous context — zero .bubble:empty, the answer painted
+# --------------------------------------------------------------------------
+
+
+def test_shared_page_has_no_empty_box(
+    page: Page,
+    browser: Browser,
+    app_url: str,
+    mock_llm: int,
+    db_ready: None,
+) -> None:
+    _reset_db(_seed)
+    page.set_default_timeout(30_000)
+
+    login(page, app_url, next="/")
+    expect(page).to_have_url(app_url + "/", timeout=30_000)
+
+    _submit(page, PLAIN_ANSWER_QUESTION)
+    _wait_settled(page)
+
+    # The phase-51 flow: the conversation is AUTO-SAVED by the time
+    # the answer settles (phase 55) — the Share click shares the
+    # linked row.
+    cookies = _admin_cookies(page)
+    _grant_clipboard(page, app_url)
+    _wait_saved_row(app_url, cookies, _auto_title(PLAIN_ANSWER_QUESTION))
+    share = page.locator("#share-chat-btn")
+    expect(share).to_be_visible()
+    _click_share_and_assert_status(page, app_url)
+    row = _find_row(_chats(app_url, cookies), _auto_title(PLAIN_ANSWER_QUESTION))
+    assert row is not None and row.get("share_url"), (
+        "the shared row must carry a share_url"
+    )
+    share_url: str = row["share_url"]
+    assert SHARE_URL_RE.fullmatch(share_url), f"bad share_url shape: {share_url}"
+    created: str = row["id"]
+    anon_ctx: BrowserContext | None = None
+    try:
+        # A FRESH context: a separate session with no cookies — the
+        # guest's only credential is the token in the URL.
+        anon_ctx = browser.new_context()
+        anon = anon_ctx.new_page()
+        anon.set_default_timeout(30_000)
+        anon.goto(app_url + share_url)
+
+        # The shared answer bubble renders normally (non-empty,
+        # painted) — and shared.js builds the identical
+        # .msg/.msg-body/.bubble structure on the same stylesheet, so
+        # ZERO child-less bubbles exist here either (locked A3).
+        bubble = anon.locator(".msg.brain .bubble").first
+        expect(bubble).to_be_visible()
+        expect(bubble).not_to_have_text("")
+        assert _bubble_height(bubble) > 0
+        assert anon.locator("#messages .msg .bubble:empty").count() == 0
+        assert anon.evaluate(PAINTED_EMPTY_BOXES_JS) == 0
+    finally:
+        if anon_ctx is not None:
+            anon_ctx.close()
+        if created:
+            _delete_chat(app_url, cookies, created)
+
+
+# --------------------------------------------------------------------------
+# 5. The empty-TEXT record corner (locked A3): a persisted brain answer
+#    of " " passes the client's length>0 filter and renders through
+#    renderMarkdown(" ") === "" — the bubble lands CHILD-LESS in the
+#    DOM and the SAME rule hides it on /?chat=<id> AND the shared page
+# --------------------------------------------------------------------------
+
+
+def test_empty_persisted_answer_paints_no_box(
+    page: Page,
+    browser: Browser,
+    app_url: str,
+    mock_llm: int,
+    db_ready: None,
+) -> None:
+    page.set_default_timeout(30_000)
+    _reset_db(_seed)
+    login(page, app_url, next="/")
+
+    # The crafted record: the whitespace brain text passes the
+    # restore's `text.length > 0` filter (a real empty string would be
+    # filtered out before rendering), and renderMarkdown(" ") is ""
+    # (unit-pinned) — so the bubble renders CHILD-LESS: exactly the
+    # :empty shape the phase-166 rule suppresses. share_token rides
+    # the same row so the shared page takes the identical corner.
+    token = uuid.uuid4()
+    with SessionLocal() as db:
+        row = SavedChat(
+            title="Phase 166 empty-answer corner",
+            messages=[
+                {"who": "user", "text": "What is the exact JSON shape?"},
+                {"who": "brain", "text": " "},
+            ],
+            share_token=token,
+        )
+        db.add(row)
+        db.commit()
+        chat_id = str(row.id)
+
+    anon_ctx: BrowserContext | None = None
+    try:
+        # The restore: the user record renders; the empty brain
+        # bubble EXISTS in the DOM (the same structure, no restore
+        # branch was added) but paints NO BOX.
+        page.goto(app_url + f"/?chat={chat_id}")
+        expect(page.locator(".msg.user .bubble").first).to_contain_text(
+            "What is the exact JSON shape?"
+        )
+        bubble = page.locator(".msg.brain .bubble").last
+        expect(bubble).to_have_count(1)
+        expect(bubble).not_to_be_visible()
+        assert _bubble_height(bubble) == 0, (
+            "the empty persisted answer must paint zero box (restore — A3)"
+        )
+        assert page.evaluate(PAINTED_EMPTY_BOXES_JS) == 0
+
+        # The shared page: the same record, a fresh anonymous context
+        # — same zero-box outcome (shared.js renders the identical
+        # structure on the same stylesheet).
+        anon_ctx = browser.new_context()
+        anon = anon_ctx.new_page()
+        anon.set_default_timeout(30_000)
+        anon.goto(app_url + f"/shared/{token}")
+        expect(anon.locator(".msg.user .bubble").first).to_contain_text(
+            "What is the exact JSON shape?"
+        )
+        anon_bubble = anon.locator(".msg.brain .bubble").last
+        expect(anon_bubble).to_have_count(1)
+        expect(anon_bubble).not_to_be_visible()
+        assert (
+            anon_bubble.evaluate("el => el.getBoundingClientRect().height") == 0
+        ), "the empty persisted answer must paint zero box (shared page — A3)"
+        assert anon.evaluate(PAINTED_EMPTY_BOXES_JS) == 0
+    finally:
+        if anon_ctx is not None:
+            anon_ctx.close()
+        with SessionLocal() as db:
+            db.execute(
+                text("DELETE FROM saved_chats WHERE id = :id"), {"id": chat_id}
+            )
+            db.commit()
@@ -270,12 +270,28 @@ def test_no_autoscroll_during_long_answer(
    # Phase 48: in-flight detection rides the Stop label (the button is
    # the enabled Stop control while a turn runs — it is no longer
    # disabled, so the disabled-state proxy is gone).
+    #
+    # WHAT IS MEASURED (2026-10-08): the VISUAL position of the text the
+    # user is reading — the first answer bubble's top relative to the
+    # viewport — not the raw scrollY. Chromium's native scroll anchoring
+    # legitimately compensates scrollY when layout above the viewport
+    # changes (at settle the Edit/Retry buttons re-target to the newest
+    # turn, so the FIRST turn's meta row collapses 44px→0): the reader
+    # sees nothing move, but a bare scrollY-constancy check would flag
+    # the compensating adjustment itself. A real up-hop (the regression
+    # this suite pins — the old block:"end" submit jump, programmatic
+    # follow-scrolls) moves the marker just the same, so the contract is
+    # asserted strictly, and correctly, on what the user perceives.
+    marker_top = """() => {
+        const el = document.querySelector('.msg.brain .bubble');
+        return el ? el.getBoundingClientRect().top : null;
+    }"""
    samples: list[float] = []
    mid_stream = 0
    deadline = time.monotonic() + 40
    while time.monotonic() < deadline:
        time.sleep(0.25)
-        samples.append(scroll_state(page)["y"])
+        samples.append(page.evaluate(marker_top))
        if page.locator("#send-label").inner_text() == "Stop":
            mid_stream += 1
        if len(samples) >= 10 and page.locator("#send-label").inner_text() == "Send":
@@ -283,7 +299,7 @@ def test_no_autoscroll_during_long_answer(
            # queued behind the stream may move the page either).
            for _ in range(3):
                time.sleep(0.3)
-                samples.append(scroll_state(page)["y"])
+                samples.append(page.evaluate(marker_top))
            break
    else:
        raise AssertionError("the long turn did not settle within 40s")
@@ -291,8 +307,8 @@ def test_no_autoscroll_during_long_answer(

    spread = max(samples) - min(samples)
    assert spread <= STABLE_PX, (
-        f"the viewport moved {spread:.1f}px while the user was scrolled up "
-        "(no-reply-autoscroll contract)"
+        f"the page moved {spread:.1f}px under the user's reading position "
+        "while they were scrolled up (no-reply-autoscroll contract)"
    )

    # The whole answer still landed (off-screen below — by design).
@@ -318,13 +334,17 @@ def test_no_autoscroll_during_thinking(
    # The thinking turn: the submit reveals the user's message (the kept
    # app scroll) ...
    submit(page, THINKING_QUESTION)
-    details = page.locator(".msg.brain").last.locator("details.thinking")
+    # 690fd98 refinement: ONE details.reasoning container per turn (the
+    # live per-run blocks are plain divs that never fold); the container
+    # is created OPEN and auto-folds on settle — same view-stability
+    # contract phase 17 pinned, one level up.
+    details = page.locator(".msg.brain").last.locator("details.reasoning")
    details.wait_for(state="attached", timeout=10_000)
    expect(details).to_have_attribute("open", "")  # created open (phase 17)
    # ... and, once the reasoning stream is clearly running ...
    page.wait_for_function(
-        "() => { const el = document.querySelector('details.thinking .thinking-text');"
-        " return !!el && el.innerText.length > 300; }",
+        "() => { const els = document.querySelectorAll('details.reasoning .thinking-text');"
+        " let n = 0; els.forEach((e) => { n += e.innerText.length; }); return n > 300; }",
        timeout=30_000,
    )
    # ... let the submit's smooth reveal settle, then the user goes up.
@@ -339,7 +359,7 @@ def test_no_autoscroll_during_thinking(
        time.sleep(0.3)
        state = page.evaluate(
            """() => {
-                const block = document.querySelector('details.thinking');
+                const block = document.querySelector('details.reasoning');
                const wrap = block ? block.closest('.msg.brain') : null;
                const el = wrap ? wrap.querySelector('.bubble') : null;
                return { y: window.scrollY,
@@ -366,7 +386,9 @@ def test_no_autoscroll_during_thinking(
    # the point of the story).
    wait_settled(page)
    assert abs(scroll_state(page)["y"]) <= STABLE_PX
-    expect(details.locator(".thinking-text")).to_contain_text(THINKING_FRAGMENT)
+    # Settled: the container auto-folded; its last run holds the tail.
+    assert not details.get_attribute("open")
+    expect(details.locator(".thinking-text").last).to_contain_text(THINKING_FRAGMENT)
    expect(page.locator(".msg.brain .bubble").last).to_contain_text(MOCK_ANSWER_MARKER)


@@ -484,9 +506,12 @@ def test_answer_content_intact(page: Page, app_url: str, seeded_kb: None) -> Non
    # (phase 17: open while streaming, closed from the first delta on).
    submit(page, THINKING_QUESTION)
    wait_settled(page)
-    details = page.locator(".msg.brain").last.locator("details.thinking")
+    # 690fd98 refinement: settled turns carry ONE details.reasoning
+    # container (auto-collapsed by settle, phase-17 contract) holding the
+    # plain per-run blocks — the per-run details.thinking is gone live.
+    details = page.locator(".msg.brain").last.locator("details.reasoning")
    expect(details).not_to_have_attribute("open")
-    expect(details.locator(".thinking-text")).to_contain_text(THINKING_FRAGMENT)
+    expect(details.locator(".thinking-text").last).to_contain_text(THINKING_FRAGMENT)
    expect(page.locator(".msg.brain .bubble").last).to_contain_text(MOCK_ANSWER_MARKER)

    # Persistence: four messages, the thinking text stored raw — and
@@ -508,9 +533,10 @@ def test_answer_content_intact(page: Page, app_url: str, seeded_kb: None) -> Non
    expect(page.locator(".msg.brain .bubble")).to_have_count(2)
    expect(page.locator(".msg.brain .bubble").first).to_contain_text(LONG_ANSWER_END)
    expect(page.locator(".msg.brain .source-chip")).to_have_count(0)
-    restored = page.locator(".msg.brain").last.locator("details.thinking")
+    # Restore writes the SAME container shape (segments-based restore).
+    restored = page.locator(".msg.brain").last.locator("details.reasoning")
    expect(restored).not_to_have_attribute("open")
-    expect(restored.locator(".thinking-text")).to_contain_text(THINKING_FRAGMENT)
+    expect(restored.locator(".thinking-text").last).to_contain_text(THINKING_FRAGMENT)
    wait_settled(page)


@@ -651,12 +651,19 @@ def test_scoped_scope_survives_save_restore(
        page.goto(app_url + f"/?chat={row['id']}")
        _assert_scoped_line(_tool_line(page))

-        # Phase 117 (D3): a restored turn renders FOLDED — one compact
-        # "Tool call(s) (N)" line (the disclosure shape is unchanged by
-        # this phase: the `open` attribute is ABSENT).
-        disclosure = page.locator(".msg.brain .tool-calls-disclosure").last
-        expect(disclosure).not_to_have_attribute("open", timeout=30_000)
-        expect(disclosure.locator("summary")).to_have_text(
+        # Phase 117 (D3), re-pinned for the phase-161 D5 refinement (the
+        # owner's 2026-10-06 shape, shipped in 690fd98): a restored turn
+        # renders FOLDED inside the ONE details.reasoning container — the
+        # container is the ONLY disclosure (zero nested <details>), the
+        # run itself is a plain div whose label is the
+        # .tool-calls-summary SPAN. The compact "Tool call (1)" line the
+        # phase-117 pin named is unchanged; only the element carrying it
+        # moved (this phase's scope — the scoped line — is asserted above).
+        container = page.locator(".msg.brain details.reasoning").last
+        expect(container).not_to_have_attribute("open", timeout=30_000)
+        expect(container.locator("details")).to_have_count(0)
+        run = page.locator(".msg.brain .tool-calls-disclosure").last
+        expect(run.locator(".tool-calls-summary")).to_have_text(
            "Tool call (1)", timeout=30_000
        )

@@ -340,12 +340,23 @@ def test_anonymous_shared_view(
        expect(bubble).to_contain_text(MOCK_ANSWER_MARKER, timeout=30_000)
        assert bubble.inner_text() == answer, "the shared answer must match the admin's"

-        # The thinking block exists and is RESTORED COLLAPSED (the
-        # phase-17 restore convention — a closed <details> has no
-        # `open` attribute).
-        think = anon.locator("details.thinking")
-        expect(think).to_have_count(1)
-        expect(think.first).not_to_have_attribute("open")
+        # The reasoning timeline exists and renders COLLAPSED (the
+        # phase-161 D5 convention, updated by the 2026-10-06 owner
+        # refinement shipped in 690fd98 — the saved turn carries
+        # `segments`, so shared.js's addSegments wraps the per-run
+        # timeline in the ONE details.reasoning container; a closed
+        # <details> has no `open` attribute, and the inner runs are
+        # plain divs — NEVER a nested <details>).
+        cont = anon.locator("details.reasoning")
+        expect(cont).to_have_count(1)
+        expect(cont.first).not_to_have_attribute("open")
+        expect(cont.locator("details")).to_have_count(0)
+        # The thinking RUN itself is present inside the collapsed
+        # container (the plain-div shape — the guest sees the same
+        # timeline the admin saw, just folded at rest).
+        runs = cont.locator("div.thinking")
+        expect(runs).to_have_count(1)
+        expect(runs.first.locator(".thinking-text")).not_to_be_empty()

        # Source chips: ZERO — the shared turn read nothing, so
        # (phase 119, LOCKED A1) its done sources are empty and the
@@ -366,9 +366,11 @@ def test_nav_switch_before_first_token_completes(
    # running — NO content frame has landed yet. (The .msg.brain bubble
    # element exists from turn start with its thinking block — the
    # pre-token state is "no content text", not "no bubble element".)
-    thinking = page.locator(".msg.brain").last.locator("details.thinking")
+    # 690fd98 refinement: the live thinking block is a plain div.thinking
+    # RUN (never folding mid-turn) inside the details.reasoning container.
+    thinking = page.locator(".msg.brain").last.locator("div.thinking")
    thinking.wait_for(state="attached", timeout=10_000)
-    expect(thinking.locator(".thinking-text")).to_contain_text(
+    expect(thinking.locator(".thinking-text").last).to_contain_text(
        THINKING_TAIL, timeout=30_000
    )
    # Still pre-token: the button is the enabled Stop control (phase 48
@@ -444,9 +446,11 @@ def test_no_orphan_brain_message_when_navigated_before_first_token(
    # which outruns Playwright's 5s assertion auto-wait on a loaded host
    # (the wait started at the FIRST thinking frame — the pre-existing
    # race, phase 55 task 02 fix).
-    thinking = page.locator(".msg.brain").last.locator("details.thinking")
+    # 690fd98 refinement: the live thinking block is a plain div.thinking
+    # RUN (never folding mid-turn) inside the details.reasoning container.
+    thinking = page.locator(".msg.brain").last.locator("div.thinking")
    thinking.wait_for(state="attached", timeout=10_000)
-    expect(thinking.locator(".thinking-text")).to_contain_text(
+    expect(thinking.locator(".thinking-text").last).to_contain_text(
        THINKING_TAIL, timeout=30_000
    )
    # Still pre-token: the button is the enabled Stop control (phase 48 —
@@ -862,7 +862,7 @@ def test_missing_folder_summaries_read_as_pending_and_self_heal(
    src_row = page.locator("#folders-tbody tr")
    expect(src_row.locator("a.folder-link")).to_have_text(SOURCE)
    expect(src_row.locator("td:nth-child(2)")).to_have_text(str(N_FILES))
-    expect(src_row.locator("td:nth-child(4) span")).to_have_text(SUM_ROOT)
+    expect(src_row.locator("td:nth-child(4) span.kb-desc-text")).to_have_text(SUM_ROOT)
    expect(page.locator(".kb-summary-pending")).to_have_count(0)

    # The gap: delete the bravo row AND the source-root row directly
@@ -877,7 +877,7 @@ def test_missing_folder_summaries_read_as_pending_and_self_heal(
    page.click("#nav-sources")
    _wait_top_level(page)
    src_row = page.locator("#folders-tbody tr")
-    s_span = src_row.locator("td:nth-child(4) span")
+    s_span = src_row.locator("td:nth-child(4) span.kb-desc-text")
    expect(s_span).to_have_text(PENDING_COPY)
    # Phase 99 (task 01): the marker toggles onto the cell's base
    # .kb-desc-text span (the one-line clamp) — the class pair, exact.
@@ -900,13 +900,13 @@ def test_missing_folder_summaries_read_as_pending_and_self_heal(
    bravo_row = rows.nth(1)
    expect(alpha_row.locator("a.folder-link")).to_have_text(FOLDER_A)
    # The intact folder: the stored line, never the marker.
-    expect(alpha_row.locator("td:nth-child(4) span")).to_have_text(SUM_ALPHA)
-    expect(alpha_row.locator("td:nth-child(4) span")).not_to_have_class(
+    expect(alpha_row.locator("td:nth-child(4) span.kb-desc-text")).to_have_text(SUM_ALPHA)
+    expect(alpha_row.locator("td:nth-child(4) span.kb-desc-text")).not_to_have_class(
        "kb-summary-pending"
    )
    # The affected folder: the marker (copy + class + title) — the
    # class pair with the phase-99 .kb-desc-text base span, exact.
-    b_span = bravo_row.locator("td:nth-child(4) span")
+    b_span = bravo_row.locator("td:nth-child(4) span.kb-desc-text")
    expect(b_span).to_have_text(PENDING_COPY)
    expect(b_span).to_have_class("kb-desc-text kb-summary-pending")
    expect(b_span).to_have_attribute("title", PENDING_TITLE)
@@ -929,7 +929,7 @@ def test_missing_folder_summaries_read_as_pending_and_self_heal(
    expect(bravo_row.locator(".kb-summary-status")).to_have_text(
        "Description updated."
    )
-    b_span = bravo_row.locator("td:nth-child(4) span")
+    b_span = bravo_row.locator("td:nth-child(4) span.kb-desc-text")
    expect(b_span).to_have_text(MANUAL_BRavo)
    expect(b_span).not_to_have_class("kb-summary-pending")

@@ -962,14 +962,14 @@ def test_missing_folder_summaries_read_as_pending_and_self_heal(
    expect(page.locator("#kb-level-title")).to_have_text(SOURCE)
    expect(page.locator("#kb-level-summary")).to_have_text(SUM_ROOT)
    rows = page.locator("#folders-tbody tr")
-    expect(rows.nth(0).locator("td:nth-child(4) span")).to_have_text(SUM_ALPHA)
-    expect(rows.nth(1).locator("td:nth-child(4) span")).to_have_text(MANUAL_BRavo)
+    expect(rows.nth(0).locator("td:nth-child(4) span.kb-desc-text")).to_have_text(SUM_ALPHA)
+    expect(rows.nth(1).locator("td:nth-child(4) span.kb-desc-text")).to_have_text(MANUAL_BRavo)
    expect(page.locator(".kb-summary-pending")).to_have_count(0)
    # And at the top level: the source row's cell carries the
    # regenerated root line, no marker.
    page.locator("#kb-crumb a.kb-crumb-link").nth(0).click()
    src_row = page.locator("#folders-tbody tr")
-    s_span = src_row.locator("td:nth-child(4) span")
+    s_span = src_row.locator("td:nth-child(4) span.kb-desc-text")
    expect(s_span).to_have_text(SUM_ROOT)
    expect(s_span).not_to_have_class("kb-summary-pending")
    expect(page.locator(".kb-summary-pending")).to_have_count(0)
@@ -36,8 +36,13 @@ thinking frame) is REPLACED by the per-run timeline (D1/D5): each
 contiguous thinking run streams into its OWN block, and the first frame
 of a new kind closes the previous run. The SSE stream is unchanged, so
 this turn renders ONE ``.tool-calls-disclosure`` + TWO ``.thinking``
-blocks (arrival order, all ABOVE the answer bubble — D2), open in their
-live windows and folded at rest. The loader pins
+blocks (arrival order) — wrapped since phase 161 in the ONE
+``details.reasoning`` container ABOVE the answer bubble (D1), the
+container open from creation until the settle fold (D2). The
+2026-10-06 D5 refinement (owner, shipped in ``690fd98``): the per-run
+blocks are PLAIN DIVS — always visible inside the container, never
+themselves collapsible (the container is the ONE disclosure). The
+loader pins
 (``test_loader_visible_from_send_through_the_tool_gap``,
 ``test_loader_a11y_and_reduced_motion``) stay unchanged — the
 never-stale loader contract (D16) is intact.
@@ -53,15 +58,17 @@ context, ``_reset_db`` reseeds the KB deterministically):
  the phase-87 elapsed counter in the tool gap, and ``#send-status``
  carries a state text (not empty) throughout.
 * ``test_thinking_runs_open_in_live_window_with_visible_loader`` —
-  THE reported symptom's state, re-pinned for phase 158 (D5 replaces
-  phase-109 D15 — a NEW block instead of a re-open): the post-delta
-  thinking frames open a per-run ``.thinking`` block (exactly one
-  block, open + non-empty ``.thinking-text`` while the bubble already
-  carries call 1's content — the run starts AFTER the first delta) with
-  the loader STILL visible (the frozen window is gone — the user sees
-  the new thinking in its own live block); the terminal state is clean
-  (loader hidden, the scratchpad holds TWO ``.thinking`` blocks —
-  EACH carrying its own round's text — BOTH folded at rest, the bubble
+  THE reported symptom's state, re-pinned for phase 158 + the phase-161
+  D5 refinement (a NEW plain-div run instead of a re-open; the runs
+  never fold, the ONE container does): while the
+  ``details.reasoning`` container is OPEN, exactly one ``.thinking``
+  run is VISIBLE with non-empty ``.thinking-text`` while the bubble
+  already carries call 1's content — the run starts AFTER the first
+  delta — with the loader STILL visible (the frozen window is gone —
+  the user sees the new thinking in its own live run); the terminal
+  state is clean (loader hidden, the container folded at the settle —
+  zero nested ``<details>`` inside it — the scratchpad holds TWO
+  ``.thinking`` runs EACH carrying its own round's text, the bubble
  carries BOTH answers, the send button reads "Send" not "Stop",
  ``#send-status`` back to the idle shape).
 * ``test_loader_a11y_and_reduced_motion`` — after a full turn: the
@@ -223,25 +230,32 @@ def test_thinking_runs_open_in_live_window_with_visible_loader(
    loader = page.locator("#turn-loader")

    # THE reported symptom's state (phase 158 D5 replaces phase-109
-    # D15 — the post-delta thinking frames open a NEW per-run block
-    # instead of re-opening the first delta's block): exactly ONE
-    # ``.thinking`` block exists, it is OPEN with non-empty text WHILE
-    # the answer bubble already carries call 1's content — the first
-    # thinking run starts AFTER the first delta (the owner's "thinking
-    # after responding" repro), and the loader is STILL visible (the
-    # frozen window is gone — the user sees the new thinking in its
-    # own live block).
+    # D15 — the post-delta thinking frames open a NEW per-run block,
+    # and the phase-161 D5 refinement (owner 2026-10-06, 690fd98) made
+    # the runs PLAIN DIVS inside the ONE details.reasoning container —
+    # always visible, the CONTAINER is the disclosure): while the
+    # container is OPEN, exactly ONE ``.thinking`` run exists and
+    # paints non-empty text WHILE the answer bubble already carries
+    # call 1's content — the first thinking run starts AFTER the first
+    # delta (the owner's "thinking after responding" repro), and the
+    # loader is STILL visible (the frozen window is gone — the user
+    # sees the new thinking in its own live run).
    page.wait_for_function(
        """() => {
-            const b = document.querySelectorAll('.msg.brain .thinking');
+            const cont = document.querySelector(
+                '.msg.brain details.reasoning[open]');
+            if (!cont) return false;
+            const b = cont.querySelectorAll('.thinking');
            const t = b.length === 1 ? b[0].querySelector('.thinking-text') : null;
-            return b.length === 1 && !!b[0].open && !!t
-                && t.textContent.trim().length > 0;
+            return b.length === 1 && !!t && t.textContent.trim().length > 0;
        }""",
        timeout=30_000,
    )
    expect(blocks).to_have_count(1)
-    expect(blocks).to_have_attribute("open", "")
+    expect(blocks.first).to_be_visible()
+    expect(
+        page.locator(".msg.brain details.reasoning").first
+    ).to_have_attribute("open", "")
    expect(blocks.nth(0).locator(".thinking-text")).to_contain_text(SENTINEL_THOUGHT)
    # The run is genuinely AFTER a delta: call 1's content is already
    # in the bubble ...
@@ -265,11 +279,12 @@ def test_thinking_runs_open_in_live_window_with_visible_loader(
    # reasoning streams — separate blocks, not one lump) ...
    expect(blocks.nth(0).locator(".thinking-text")).to_contain_text(SENTINEL_THOUGHT)
    expect(blocks.nth(1).locator(".thinking-text")).to_contain_text(SENTINEL_FINAL_THOUGHT)
-    # ... and BOTH runs are FOLDED at rest (D5 — the per-run
-    # fold-at-rest; the phase-109 symptom — the frozen chat — stays
-    # fixed: a run is open in its live window, never re-opened after
-    # done) ...
-    expect(page.locator(".msg.brain .thinking[open]")).to_have_count(0)
+    # ... and the ONE container FOLDED at the settle (phase 161 D2 —
+    # the container is the ONLY disclosure; the inner runs are plain
+    # divs, never collapsible: zero nested <details>; the phase-109
+    # symptom — the frozen chat — stays fixed) ...
+    expect(page.locator(".msg.brain details.reasoning[open]")).to_have_count(0)
+    expect(page.locator(".msg.brain details.reasoning details")).to_have_count(0)
    # ... the bubble carries BOTH answers (call 1 + call 2) ...
    expect(bubble).to_contain_text(SENTINEL_FIRST)
    expect(bubble).to_contain_text(SENTINEL_FINAL)
@@ -0,0 +1,628 @@
+"""Unit: the empty answer bubble never paints (phase 166, task 01,
+``TODO.md`` L4 — "There's an empty box where the LLM response will go
+that shows up before the LLM responds. It looks ugly." (owner's
+``image.png``: Thinking / Tool calls (2) / an expanded Thinking run,
+then a bordered, padded, EMPTY rounded box)).
+
+Why the box exists: the first ``thinking``/``tool`` frame of a turn
+calls ``addMessage("brain", "")`` (app.js ~4063 / ~4164 / ~4269) — the
+EMPTY bubble is created ON PURPOSE as the DOM anchor for
+``ensureReasoningWrap`` (``body.insertBefore(cont,
+body.querySelector(".bubble"))`` — the phase-158/161 per-run timeline
+sits ABOVE the answer bubble). ``.bubble`` carries its own padding +
+border + background + box-shadow, so an empty one paints the reported
+~40px box (and ``.msg-body``'s flex gap adds the separation).
+
+The fix is CSS-only (LOCKED A1 — owner 2026-07-10: "hide until the
+first token", NOT "restructure the anchor"):
+``.msg .bubble:empty { display: none; }`` in the Messages block.
+``:empty`` matches only a child-less element — exactly
+``innerHTML === ""`` — and the first delta writes content, so the box
+appears AT the first token; ``display: none`` also drops the flex item
+so the gap collapses. Because ``renderMarkdown("")`` returns ``""``
+(node-executed below), the restore path and the shared page (same
+``.msg/.msg-body/.bubble`` structure, same stylesheet — locked A3) are
+covered by the same rule. The typing indicator is never ``:empty``
+(three dot ``<span>``s), so the loading state never disappears (locked
+A2 — nothing replaces the box).
+
+Browser behavior is E2E-covered by ``tests/e2e/test_no_empty_box.py``
+(task 03); here we pin the source-level contract in the house
+"read-the-source" style (test_thinking_scroll.py /
+test_markdown_tables.py): the exact declaration exists ONCE in the
+Messages block, ``.bubble`` keeps its paint (the rule HIDES, never
+restyles), the typing-dot rules are untouched, the JS anchors survive
+unchanged (D2 — this is the pin that stops a future "cleanup" from
+moving the fix into JS and breaking the timeline), and the empty-input
+renderer behavior is executed under node.
+
+AUDIT NOTE (phase 166 task 02 — the §7.4 "never stale" inventory; the
+record the next agent reads instead of re-deriving it). With the empty
+bubble suppressed, EVERY mid-turn state still has a visible
+affordance, in order:
+  • idle / pre-thinking → ``#typing-indicator``: the ``role="status"``
+    bubble labelled ``TYPING_LABEL()`` with its three dot ``<span>``s
+    (never ``:empty``, so the rule can never hide it);
+  • long tool call / frameless silence → the ``thinkingClock`` still
+    paints the ``Ns`` ``.typing-elapsed`` suffix INTO that same bubble
+    (a child ⇒ the bubble is never ``:empty`` ⇒ the clock keeps
+    painting);
+  • thinking / tool → the ``details.reasoning`` container
+    (``ensureReasoningWrap`` sets ``cont.open = true`` — visible while
+    live) with its live status span: "Thinking…" / "Calling <name>…"
+    / the settle "N steps";
+  • first token → the answer bubble APPEARS: the first ``delta``
+    writes ``innerHTML``, so ``:empty`` stops matching on the SAME
+    frame the box returns — and ``#messages`` (the ONE conversation
+    ``aria-live="polite"`` region) announces as content lands, at
+    which point the bubble is visible again anyway;
+  • terminal → the settle copy: answer / deflection (.maybe-try chips)
+    / stopped (.stopped-note) / failed (.failed-note) + the
+    ``#send-status`` / ``#send-label`` composer signals.
+Nothing measures the hidden box (``display: none`` yields zero
+rects), so no scroll or pin can shift: ``scrollReveal`` receives the
+``.msg`` WRAP (never the bubble), the phase-161 window pins take
+``.reasoning-body``, and ``.bubble`` is only ever read for CONTENT
+(``textContent`` / ``innerHTML`` / ``querySelector`` / ``classList`` /
+``prepend`` / ``setAttribute``) or used as an ``insertBefore`` anchor.
+The ``.msg-meta`` rows (chips, figures, actions, notes) append to
+``.msg-body`` as SIBLINGS of the bubble (phase 130), so hiding an
+empty one never hides them. The restore
+(``addMessage("brain", renderMarkdown(m.text), true)``) and the shared
+page (``shared.js`` builds the identical ``.msg/.msg-body/.bubble``
+structure on the same stylesheet) land in the same ``.msg `` scope —
+no ``shared.js`` change needed (locked A3).
+"""
+from __future__ import annotations
+
+import json
+import re
+import shutil
+import subprocess
+from pathlib import Path
+
+import pytest
+
+FRONTEND = Path(__file__).resolve().parents[2] / "frontend"
+ASSETS = FRONTEND / "assets"
+APP_JS = ASSETS / "app.js"
+MARKDOWN_JS = ASSETS / "markdown.js"
+SHARED_JS = ASSETS / "shared.js"
+STYLES_CSS = ASSETS / "styles.css"
+INDEX_HTML = FRONTEND / "index.html"
+
+RULE = ".msg .bubble:empty { display: none; }"
+
+HAVE_NODE = shutil.which("node") is not None
+
+
+def _text(path: Path) -> str:
+    assert path.is_file(), f"missing frontend file: {path}"
+    return path.read_text(encoding="utf-8")
+
+
+def _css() -> str:
+    return _text(STYLES_CSS)
+
+
+def _js() -> str:
+    return _text(APP_JS)
+
+
+def _bubble_rule(css: str) -> str:
+    """Body of the bare `.bubble { ... }` rule (the answer bubble)."""
+    rule = re.search(r"\n\.bubble \{([\s\S]*?)\n\}", css)
+    assert rule, "styles.css must style .bubble"
+    return rule.group(1)
+
+
+def _function_body(js: str, signature: str) -> str:
+    """Source of one top-level function, from its signature to the
+    first column-0 `}` (the house indentation makes that the function
+    end). Used by the task-02 audit pins."""
+    start = js.find(signature)
+    assert start != -1, f"missing function: {signature}"
+    end = js.find("\n}", start)
+    assert end != -1, f"unterminated function: {signature}"
+    return js[start:end]
+
+
+def _strip_comments(src: str) -> str:
+    """Crude-but-sufficient comment stripper for the audit greps: block
+    comments then line comments (the house files carry no `//` inside
+    string literals — no-CDN pins keep URLs out)."""
+    src = re.sub(r"/\*[\s\S]*?\*/", "", src)
+    return re.sub(r"//[^\n]*", "", src)
+
+
+def _run_node(script: str) -> str:
+    proc = subprocess.run(["node", "-e", script], capture_output=True, text=True, timeout=60)
+    assert proc.returncode == 0, f"node failed: {proc.stderr}"
+    return proc.stdout
+
+
+# ---------- styles.css: the one rule ----------
+
+
+def test_suppression_rule_exists_exactly_once() -> None:
+    """The exact declaration ``.msg .bubble:empty { display: none; }``
+    exists in styles.css EXACTLY once — one rule, no duplicates, no
+    near-miss variant (a `.bubble:empty` without the `.msg` scope would
+    leak past the Messages section)."""
+    css = _css()
+    assert css.count(RULE) == 1, "the suppression rule must exist exactly once"
+    assert css.count(".bubble:empty") == 1, "no unscoped or extra :empty variant"
+
+
+def test_rule_sits_in_the_messages_block() -> None:
+    """Provenance ordering (the test_date_editor.py css.find idiom):
+    the rule lives in the Messages block, directly after the
+    ``.bubble p`` content rule and before ``.bubble pre`` — i.e. after
+    ``.bubble {`` — so the fix reads next to the bubble it hides."""
+    css = _css()
+    assert css.find("---------- Messages ----------") < css.find(RULE), (
+        "the rule must sit in the Messages block"
+    )
+    assert css.find(".bubble {") < css.find(".bubble p {") < css.find(RULE), (
+        "the rule must follow the .bubble rules it suppresses"
+    )
+    assert css.find(RULE) < css.find(".bubble pre {"), (
+        "the rule must stay between .bubble p and .bubble pre"
+    )
+
+
+def test_rule_is_scoped_to_msg_bubbles() -> None:
+    """The selector is `.msg .bubble:empty` — the chat/shared-page
+    message bubbles ONLY. The document viewer's .doc-* blocks, the
+    composer, and every other :empty-carrying element never match."""
+    css = _css()
+    assert RULE in css
+    # Every `.bubble:empty` occurrence carries the `.msg ` prefix — no
+    # bare variant escaping the Messages scope — and there is no
+    # broader `.msg:empty` (which would hide the whole message wrap).
+    for m in re.finditer(r"\.bubble:empty", css):
+        assert css[: m.start()].endswith(".msg "), (
+            "the :empty rule must stay scoped to .msg .bubble"
+        )
+    assert ".msg:empty" not in css
+
+
+def test_bubble_keeps_its_paint_the_rule_hides_never_restyles() -> None:
+    """The bare `.bubble` rule still declares padding, border-radius,
+    background, border and box-shadow (LOCKED A1: the fix HIDES the
+    empty box, it never restyles a non-empty bubble) — and no
+    `display: none` crept into the base rule itself."""
+    body = _bubble_rule(_css())
+    for decl in ("padding", "border-radius", "background", "border", "box-shadow"):
+        assert decl in body, f".bubble must keep its {decl} (hide, not restyle)"
+    assert "display: none" not in body, "the base .bubble rule must stay visible"
+
+
+def test_suppression_carries_phase_166_provenance() -> None:
+    """The provenance comment directly above the rule names the defect
+    (TODO L4 + the owner's image.png), the JS anchors it preserves
+    (addMessage("brain","") / ensureReasoningWrap), and why the typing
+    indicator is unaffected."""
+    css = _css()
+    idx = css.find(RULE)
+    comment = css[:idx]
+    start = comment.rindex("/*")
+    note = css[start:idx]
+    assert "Phase 166" in note
+    assert "TODO L4" in note, "the comment must name the TODO line"
+    assert "image.png" in note, "the comment must name the owner's screenshot"
+    assert 'addMessage("brain","")' in note or 'addMessage("brain", "")' in note, (
+        "the comment must name the empty-bubble creation it hides"
+    )
+    assert "ensureReasoningWrap" in note, "the comment must name the preserved anchor"
+    assert "typing" in note, "the comment must explain why the indicator is safe"
+
+
+def test_typing_indicator_rules_are_untouched() -> None:
+    """The loading affordance stays painted: the `.typing` bubble rule
+    (inline-flex + gap) and the three-dot animation rules
+    (`.typing span`, the two animation-delays, @keyframes typing) are
+    all still present — the indicator bubble carries the dot children,
+    so the new :empty rule can never hide it (locked A2)."""
+    css = _css()
+    assert ".typing { display: inline-flex" in css
+    assert "animation: typing 1.2s infinite ease-in-out;" in css
+    assert ".typing span:nth-child(2) { animation-delay: 0.15s; }" in css
+    assert ".typing span:nth-child(3) { animation-delay: 0.3s; }" in css
+    assert "@keyframes typing {" in css
+
+
+# ---------- app.js: the CSS-only contract (D2) ----------
+
+
+def test_addtyping_bubble_is_never_empty() -> None:
+    """`addTyping()` still builds the bubble with its three dot
+    `<span>` children — `:empty` matches only a CHILD-LESS element, so
+    the typing indicator (role=status) is provably never hidden by the
+    new rule: the loading state cannot disappear."""
+    js = _js()
+    start = js.find("function addTyping() {")
+    assert start != -1, "app.js must keep addTyping()"
+    body = js[start : js.find("\n}", start)]
+    assert '<div class="bubble typing"' in body
+    assert "<span></span><span></span><span></span>" in body, (
+        "the three dot spans are the non-:empty guarantee"
+    )
+
+
+def test_the_three_empty_bubble_anchors_survive() -> None:
+    """D2 pin: app.js is NOT modified by this phase — all three
+    `addMessage("brain", "")` call sites (thinking ~4063 / tool ~4164 /
+    first delta ~4269) still create the empty wrap. A future "cleanup"
+    that moves the fix into JS would re-own the phase-158/161 anchors
+    for zero user-visible gain over the CSS rule — this pin stops it."""
+    js = _js()
+    assert js.count('addMessage("brain", "")') == 3, (
+        "the thinking/tool/first-delta empty-bubble anchors must all stay"
+    )
+
+
+def test_ensure_reasoning_wrap_still_anchors_before_the_bubble() -> None:
+    """`ensureReasoningWrap` still does `body.insertBefore(cont,
+    body.querySelector(".bubble"))` — the hidden bubble remains the
+    DOM anchor the per-run timeline inserts BEFORE (the CSS rule hides
+    the box; it never removes the element)."""
+    js = _js()
+    start = js.find("function ensureReasoningWrap(wrap) {")
+    assert start != -1, "app.js must keep ensureReasoningWrap"
+    body = js[start : js.find("\n}", start)]
+    assert 'body.insertBefore(cont, body.querySelector(".bubble"))' in body
+
+
+# ---------- markdown.js: WHY one rule covers restore + shared (A3) ----------
+
+
+@pytest.mark.skipif(not HAVE_NODE, reason="node not available")
+def test_rendermarkdown_empty_inputs_yield_empty_string() -> None:
+    """Executed under node (the phase-44 driver idiom):
+    `renderMarkdown("")` and `renderMarkdown("   ")` BOTH return `""`
+    — so a persisted answer with empty/whitespace text lands as a truly
+    child-less `<div class="bubble"></div>` in the restore AND the
+    shared page, and the same `:empty` rule covers them (locked A3)."""
+    js = _text(MARKDOWN_JS)
+    driver = (
+        js
+        + "\nconst out = { empty: renderMarkdown(\"\"), blank: renderMarkdown(\"   \") };\n"
+        + "console.log(JSON.stringify(out));"
+    )
+    results = json.loads(_run_node(driver))
+    assert results["empty"] == "", f'renderMarkdown("") must be "", got {results["empty"]!r}'
+    assert results["blank"] == "", f'renderMarkdown("   ") must be "", got {results["blank"]!r}'
+
+
+# ---------- no-CDN (AGENTS.md rule 6) ----------
+
+
+def test_no_cdn_introduced() -> None:
+    """The touched assets stay CDN-free (AGENTS.md rule 6): no remote
+    url()/@import in styles.css, no remote <script src> in app.js."""
+    css = _css()
+    assert "url(http" not in css and "@import url(" not in css, (
+        "no CDN (AGENTS.md rule 6)"
+    )
+    js = _js()
+    assert "src=\"http" not in js and "src='http" not in js
+
+
+# ---------- §7.4 feedback inventory (task 02) ----------
+#
+# Each test names the mid-turn state whose affordance it guarantees.
+# If one of these pins fails, a turn has become feedback-less — the
+# exact regression phase 166 must not ship (PLAN §7.4, AGENTS.md 10).
+
+
+def test_pre_thinking_state_typing_indicator_is_a_labeled_status() -> None:
+    """STATE: idle → pre-thinking. `addTyping()` still builds the
+    #typing-indicator bubble as a `role="status"` live element with the
+    accessible label `aria-label="${TYPING_LABEL()}"` and the three
+    `<span></span>` dots — the dots are both the visible pulse and the
+    non-`:empty` guarantee (task 01's pin), so this affordance survives
+    the suppression rule untouched (locked A2)."""
+    body = _function_body(_js(), "function addTyping() {")
+    assert 'wrap.id = "typing-indicator"' in body
+    assert 'role="status"' in body
+    assert 'aria-label="${TYPING_LABEL()}"' in body
+    assert body.count("<span></span>") == 3, (
+        "the three dot spans are the visible pre-thinking affordance"
+    )
+
+
+def test_pre_thinking_state_setuistate_arms_typing_and_clock() -> None:
+    """STATE: idle → thinking. `setUiState` still arms BOTH composer-
+    side and chat-side feedback on the thinking transition —
+    `addTyping()` + `startThinkingClock()` — and the persistent turn
+    loader stays owned by the single `turnLoader.hidden = !inFlight`
+    line (phase 109 D16). The suppression rule lives nowhere near
+    this state machine."""
+    body = _function_body(_js(), 'export function setUiState(state, errorDetail = "", opts = {}) {')
+    assert "addTyping();" in body
+    assert "startThinkingClock();" in body
+    assert "turnLoader.hidden = !inFlight;" in body
+    assert 'sendBtn.classList.toggle("is-stop", inFlight);' in body
+
+
+def test_long_tool_call_state_clock_paints_the_elapsed_suffix() -> None:
+    """STATE: long tool call (frameless silence ≥ 10s). `thinkingClock`
+    still writes the `Ns` suffix into `#typing-indicator .bubble` as a
+    `.typing-elapsed` child (`el.textContent = secs + "s"`) and keeps
+    the aria channel byte-identical (`is still thinking (${secs}s)`).
+    NOTE: the suffix APPENDS a child to the typing bubble — that bubble
+    is therefore NEVER `:empty` (dots from creation, the suffix on top)
+    — so the clock keeps painting for the whole turn; the suppression
+    rule cannot starve this affordance."""
+    body = _function_body(_js(), "function startThinkingClock() {")
+    assert 'document.querySelector("#typing-indicator .bubble")' in body
+    assert 'el.className = "typing-elapsed"' in body
+    assert 'el.textContent = secs + "s"' in body
+    assert "is still thinking (${secs}s)" in body
+    assert "bubble.appendChild(el)" in body
+
+
+def test_thinking_tool_state_reasoning_status_literals_stay_wired() -> None:
+    """STATE: thinking / tool / the settle. The D6 status literals are
+    still wired at their call sites: `setReasoningStatus("Thinking…")`
+    on every thinking frame (1 site), `setReasoningStatus(
+    `Calling ${name}…`)` on every tool frame (1 site), and the settle
+    `${segRuns.length} steps` at its THREE settle sites (done, the
+    user stop, the failed-turn funnel in finalizeFailedTurn). The live
+    status span is the text signal of the thinking/tool state while
+    the answer box is still hidden."""
+    js = _js()
+    assert js.count('setReasoningStatus("Thinking…")') == 1
+    assert js.count("setReasoningStatus(`Calling ${name}…`)") == 1
+    assert js.count("setReasoningStatus(`${segRuns.length} steps`)") == 3, (
+        "done + stop + failed-settle must all keep the N-steps literal"
+    )
+
+
+def test_thinking_tool_state_reasoning_container_is_visible_while_live() -> None:
+    """STATE: thinking / tool. `ensureReasoningWrap` still sets
+    `cont.open = true` — the per-run timeline container is OPEN (i.e.
+    actually VISIBLE) from the first thinking/tool frame until
+    settleReasoning folds it (phase 161 D1/D2). With the answer box
+    hidden, this container is the turn's main brain-side surface; a
+    silently-closed container would break §7.4 even though the box
+    fix itself is CSS-only."""
+    body = _function_body(_js(), "function ensureReasoningWrap(wrap) {")
+    assert "cont.open = true;" in body
+
+
+def test_composer_side_state_copy_is_untouched() -> None:
+    """STATE: every non-idle state, composer side. The `#send-label`
+    ("Send" ⇄ "Stop") and the `#send-status` live region
+    (`SEND_STATUS[state]` — "… is thinking" / "… is answering" /
+    "The last question failed — try again") are still written by
+    `setUiState` exactly as before — the phase-166 change lives in one
+    CSS rule and never touches the composer-side signal (locked A2:
+    nothing replaces the box because the existing states already
+    carry the feedback)."""
+    js = _js()
+    start = js.find("const SEND_STATUS = Object.freeze({")
+    assert start != -1
+    block = js[start : js.find("});", start)]
+    assert "is thinking" in block and "is answering" in block
+    assert "The last question failed — try again" in block
+    body = _function_body(js, 'export function setUiState(state, errorDetail = "", opts = {}) {')
+    assert 'sendLabel.textContent = inFlight ? "Stop" : "Send";' in body
+    assert 'sendStatus.textContent = SEND_STATUS[state]?.() ?? "";' in body
+
+
+def test_messages_stays_the_one_conversation_live_region() -> None:
+    """STATE: first token → the announcement channel. `#messages`
+    keeps `aria-live="polite"` (frontend/index.html — the ONE
+    conversation live region; the composer-side #send-status and the
+    per-feature announcers are separate surfaces). An aria-live
+    announcement fires when the answer CONTENT lands in the region —
+    and by construction that is the first delta, the same frame the
+    bubble gains children and becomes visible again: the suppression
+    rule can therefore never hide content from the a11y channel."""
+    html = _text(INDEX_HTML)
+    m = re.search(r'<section class="messages"[^>]*>', html)
+    assert m, "index.html must keep the #messages section"
+    tag = m.group(0)
+    assert 'id="messages"' in tag
+    assert 'aria-live="polite"' in tag, (
+        "the conversation region stays the live region (PLAN §7.4)"
+    )
+
+
+# ---------- nothing measures the hidden box (task 02) ----------
+
+MEASURE_PROPS = (
+    "getBoundingClientRect",
+    "offsetHeight",
+    "offsetWidth",
+    "clientHeight",
+    "clientWidth",
+    "scrollHeight",
+    "scrollTop",
+    "scrollLeft",
+    "offsetTop",
+    "offsetLeft",
+)
+
+
+def test_scroll_reveal_receives_the_msg_wrap_not_the_bubble() -> None:
+    """FINDING: `addMessage` calls `if (scroll) scrollReveal(wrap);`
+    with the `.msg` WRAP — never the `.bubble` — and `scrollReveal`
+    measures NOTHING of its argument (`void wrap;` then a document-
+    bottom `window.scrollTo`). `display: none` on the bubble therefore
+    cannot shift a scroll: the only page scroll paths are the two
+    user-intent call sites, and none of them touches the hidden box."""
+    add = _function_body(_js(), "function addMessage(who, html, scroll = false, image = null) {")
+    assert "if (scroll) scrollReveal(wrap);" in add
+    assert 'wrap.className = `msg ${who}`' in add, "wrap is the .msg element"
+    assert 'querySelector(".bubble")' not in add.split("scrollReveal")[-1], (
+        "the scroll call must not hand the bubble to scrollReveal"
+    )
+    reveal = _function_body(_js(), "function scrollReveal(wrap, behavior = SCROLL) {")
+    assert "void wrap;" in reveal
+    assert "window.scrollTo({ top: document.documentElement.scrollHeight, behavior });" in reveal
+    assert ".bubble" not in reveal
+
+
+def test_reasoning_window_pins_never_take_a_bubble() -> None:
+    """FINDING: the phase-161 window pins operate on `.reasoning-body`
+    handles (`reasoningWrapEl?.querySelector(".reasoning-body")`, both
+    frame sites) — `isReasoningPinned` / `scrollReasoningToBottom`
+    take the container window `el`, NEVER a `.bubble`. The pinned-tail
+    follow is therefore blind to the bubble's box: hiding it cannot
+    move a pin."""
+    js = _js()
+    pinned = _function_body(js, "function isReasoningPinned(el) {")
+    scroll = _function_body(js, "function scrollReasoningToBottom(el) {")
+    assert ".bubble" not in pinned and ".bubble" not in scroll
+    assert js.count('reasoningWrapEl?.querySelector(".reasoning-body")') == 2, (
+        "the thinking and tool frame sites capture the CONTAINER window"
+    )
+    assert "isReasoningPinned(rbody)" in js and "isReasoningPinned(tbody)" in js
+    assert "scrollReasoningToBottom(rbody)" in js and "scrollReasoningToBottom(tbody)" in js
+
+
+def test_no_line_measures_a_bubble_box() -> None:
+    """FINDING (grep audit): NO line of app.js or shared.js pairs a
+    geometry property (getBoundingClientRect / offsetHeight /
+    clientHeight / scroll* / offset*) with `bubble`. The only geometry
+    reads that exist today are pinned to their OWN surfaces: the
+    thinking-text window (`textEl`), the container window (`el`, the
+    two pins above), the composer textarea autosize (`input`), and the
+    document-bottom scroll (documentElement) — so `display: none` on
+    an empty `.bubble` cannot shift a scroll or a pin ANYWHERE."""
+    for path in (APP_JS, SHARED_JS):
+        for line in _text(path).splitlines():
+            if any(prop in line for prop in MEASURE_PROPS):
+                assert "bubble" not in line.lower(), (
+                    f"{path.name} measures a box on a bubble line: {line.strip()}"
+                )
+
+
+def test_bubble_selector_is_only_lookups_and_anchors() -> None:
+    """FINDING: every code line (comments stripped) that references the
+    `.bubble` SELECTOR is a `querySelector` lookup or an
+    `insertBefore` anchor — the two operations `display: none` does
+    not disturb (the element stays in the DOM; locked A1/D2). No
+    measurement ever hangs off the selector itself."""
+    for path in (APP_JS, SHARED_JS):
+        code = _strip_comments(_text(path))
+        selector_lines = [
+            ln.strip()
+            for ln in code.splitlines()
+            if re.search(r"\.bubble(?=[\"'])", ln)
+        ]
+        assert selector_lines, f"{path.name} must still reference .bubble"
+        for ln in selector_lines:
+            assert "querySelector" in ln, (
+                f"the .bubble selector may only appear in lookups/anchors: {ln}"
+            )
+
+
+def test_bubble_variable_is_only_touched_for_content() -> None:
+    """FINDING: the `bubble` variable (whatever a lookup bound) is
+    only ever touched for CONTENT — textContent, innerHTML,
+    classList, querySelector, prepend, appendChild, setAttribute — in
+    BOTH app.js and shared.js. A geometry property on the handle
+    would read ZERO once the box is hidden, so the allowlist failing
+    here means some future code silently depends on the painted box —
+    exactly what this audit forbids."""
+    content_props = {
+        "textContent",
+        "innerHTML",
+        "classList",
+        "querySelector",
+        "prepend",
+        "appendChild",
+        "setAttribute",
+    }
+    for path in (APP_JS, SHARED_JS):
+        code = _strip_comments(_text(path))
+        touched = set(re.findall(r"(?<![.\w])bubble\??\.\s*([A-Za-z_$][\w$]*)", code))
+        assert touched <= content_props, (
+            f"{path.name} touches non-content props on a bubble: "
+            f"{sorted(touched - content_props)}"
+        )
+    # The documented content reads exist (a vacuous allowlist is not
+    # the contract — these are the sites the audit enumerated):
+    js = _strip_comments(_js())
+    assert 'bubble.textContent' in js
+    assert 'bubble.querySelector(".msg-image")' in js
+    assert 'bubble.classList.remove("is-editing")' in js
+    assert 'wrap.querySelector(".bubble").innerHTML = renderMarkdown' in _js()
+
+
+# ---------- .msg-meta rows stay OUTSIDE the bubble (task 02) ----------
+
+
+def test_meta_rows_append_to_msg_body_never_into_the_bubble() -> None:
+    """Phase-130 pin: the chips row, the images row, the answer figure
+    and the actions row all append to `.msg-body` — SIBLINGS of
+    `.bubble` — and none of the appender bodies even mentions
+    `.bubble`. Hiding an empty bubble therefore never hides a chip,
+    an action, or a figure; a future refactor that moves them INTO the
+    bubble would fail this pin (and the owner's screenshot defect
+    would return multiplied)."""
+    js = _js()
+    actions = _function_body(js, "function getActionsRow(wrap) {")
+    assert 'row.className = "msg-meta msg-actions";' in actions
+    assert "body.appendChild(row);" in actions
+    assert ".bubble" not in actions
+    sources = _function_body(js, "function appendSources(wrap, sources) {")
+    assert 'const body = wrap.querySelector(".msg-body");' in sources
+    assert 'meta.className = "msg-meta";' in sources
+    assert "body.appendChild(meta);" in sources, "chips row → .msg-body"
+    assert "body.appendChild(images);" in sources, "images row → .msg-body"
+    assert ".bubble" not in sources
+    figure = _function_body(js, "function appendAnswerImage(wrap, ref) {")
+    assert 'const body = wrap.querySelector(".msg-body");' in figure
+    assert "body.appendChild(fig);" in figure, "answer figure → .msg-body"
+    assert ".bubble" not in figure
+    for name, cls in (("appendStoppedNote(wrap)", "stopped-note"),
+                      ("appendFailedNote(wrap, detail)", "failed-note")):
+        note = _function_body(js, f"function {name} {{")
+        assert "const meta = getActionsRow(wrap);" in note, (
+            f".{cls} joins the actions row (itself a .msg-body child)"
+        )
+        assert f'note.className = "{cls}";' in note
+        assert "meta.appendChild(note);" in note
+        assert ".bubble" not in note
+
+
+# ---------- the restore and shared corners (task 02) ----------
+
+
+def test_restore_renders_brain_text_through_the_pinned_call_shape() -> None:
+    """CORNER: the restore. The phase-14 restore renders every brain
+    record through `addMessage("brain", renderMarkdown(m.text), true)`
+    — with `m.text === ""` the node-pinned `renderMarkdown("") === ""
+    ` (pinned above) makes the bubble CHILD-LESS ⇒ `:empty` ⇒ hidden:
+    an empty persisted answer paints no box, through the SAME rule
+    (locked A3) — no restore-path branch was added."""
+    js = _js()
+    assert js.count('addMessage("brain", renderMarkdown(m.text), true)') == 1
+
+
+def test_shared_page_builds_the_identical_msg_structure_no_change_needed() -> None:
+    """CORNER: the shared page. `addSharedMessage` builds the same
+    `wrap.className = `msg ${who}`` wrap and the identical
+    `<div class="bubble">${html}</div>` structure (shared.js ~113) as
+    `addMessage` — so the rule's `.msg ` scope prefix matches BOTH
+    pages (shared.html loads the same styles.css). shared.js carries
+    NO `:empty` rule of its own: one stylesheet, one rule, no
+    shared.js change needed (locked A3)."""
+    shared = _text(SHARED_JS)
+    body = _function_body(shared, "function addSharedMessage(who, html) {")
+    assert "wrap.className = `msg ${who}`" in body
+    assert '<div class="bubble">${html}</div>' in body
+    # The chat page builds the byte-identical bubble fragment:
+    assert '<div class="bubble">${html}</div>' in _js()
+    # The rule's scope prefix is `.msg ` — the class both pages put
+    # on the wrap — and shared.js adds no second suppression rule.
+    assert RULE.startswith(".msg .bubble:empty")
+    assert ":empty" not in shared
+    # shared.html is on the same stylesheet (one rule covers it).
+    shared_html = _text(FRONTEND / "shared.html")
+    assert 'href="/assets/styles.css"' in shared_html