Six-page app: sidebar shell with Juniper's portrait, whole-workflow IA

The two-page dashboard/review split becomes a proper information
architecture: Pipeline (stages + live activity), Photos (drag-and-drop,
gallery with per-photo extraction state, reshoot tickets — photo work
lives with photos), Review (decisions only, keyboard-first), Catalog
(the full title ledger with filtering), Queue (what upload will do and
everything it has done), and Library (the enriched collection browser,
with an honest empty state until real BGG data lands). Pages render
server-side from a shared shell — sidebar rail with the rainbow path
running its edge, live count badges on Photos/Review/Queue, and
Juniper's full portrait finally displayed, with her credit and a
standard third-party trademark attribution beneath it (one notice, not
per-mention symbols — the convention for referring to another party's
mark).

Shared client plumbing moves to static/app.js (escaping contract
documented at the innerHTML sink). New endpoints: /api/photos-list,
/api/queue, /api/library, plus a reshoot count in /api/pipeline.
Screenshot review caught two real bugs: photos-list crashed on
bare-array raw caches, and .DS_Store was listed as a shelf photo —
photo_names() now filters by suffix everywhere, including the /photos
allowlist.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

CLAUDE.md

@@ -20,14 +20,14 @@ Full design lives in `bgg-shelf-pipeline-spec.md` (read it before changing pipel
 ## Commands
 
 - `uv sync` — install deps (Python 3.12+, managed by **uv**; use `uv add`, never pip). `uv run bggpipe init` handles first-run setup (folders, .env credentials, the one-time `playwright install chromium`).
-- `uv run bggpipe web` — the full pipeline as a local web app (dashboard at `/`, review at `/review`); stage runs execute one-at-a-time in a background job.
+- `uv run bggpipe web` — the app: six pages (Pipeline `/`, Photos, Review, Catalog, Queue, Library) in a shared sidebar shell; stage runs execute one-at-a-time in a background job.
 - `uv run bggpipe <stage>` — run a pipeline stage. Non-secret settings come from `config.toml` (username, dirs, vision model, rate limit); `--config` overrides the path.
 - `uv run pytest` — the suite runs fully offline against fixtures. Tests marked `live` hit the real BGG API (read-only) and are skipped unless you pass `--run-live`.
 - `uv run ruff check` / `uv run ruff format` — lint (rules E, F, I, UP, B, SIM) and format.
 
 ## Layout
 
-- `src/bggpipe/` — `cli.py` (typer app), one module per stage (`extract`, `resolve`, `review` + `webreview`, `diff`, `upload`, `enrich`); `templates/` + `static/app.css` are the web UI (one shared stylesheet is the design system — its tokens derive from the mascot art), plus `bgg_client.py` (rate-limited XML API2 client that caches responses to `data/bgg_cache/`), `jobs.py` (single-slot background stage runner for the web UI), `normalize.py` (title normalization), `models.py` (dataclasses), `config.py`, `fsio.py` (atomic writes), `init_wizard.py` (first-run setup).
+- `src/bggpipe/` — `cli.py` (typer app), one module per stage (`extract`, `resolve`, `review` + `webreview`, `diff`, `upload`, `enrich`); `templates/shell.html` + `templates/pages/*` + `static/app.{css,js}` are the web UI (the stylesheet is the design system — tokens derive from the mascot art), plus `bgg_client.py` (rate-limited XML API2 client that caches responses to `data/bgg_cache/`), `jobs.py` (single-slot background stage runner for the web UI), `normalize.py` (title normalization), `models.py` (dataclasses), `config.py`, `fsio.py` (atomic writes), `init_wizard.py` (first-run setup).
 - `scripts/` — `write_stub_fixtures.py` / `write_photo_fixtures.py` generate synthetic fixtures; `record_fixtures.py` re-records real API responses once a token exists.
 - `tests/fixtures/bgg_cache/` — stub XML fixtures the offline tests run against.
 - `data/` — pipeline state (CSV/JSON artifacts are committed; caches are not — see Git).