chore(archive): park SWE-bench Benchmark (Beta) UI in .archive

[Harry19081]

Stacked on #821 (dev/dissolve-hooks-workstation). Merge #821 first, then retarget this PR to develop (GitHub does it automatically when the base branch is deleted after merge).

Summary

• Park the SWE-bench "Benchmark (Beta)" UI in .archive/ following the repo's existing convention (paths mirror src/, excluded from tsconfig, restorable by reverse git mv; full map appended to .archive/README.md).
• Sever the feature's branches in shared code: benchmark WorkStation tab type, the chat-panel BENCHMARK create target + benchmarkSessionGroup surface, the nav-sidebar coordinator-session routing / child-session hiding, the flask-conical session icon override, E2E bootstrap wiring, wdio docker fixture, and creator.benchmark.* i18n keys in 13 locales.
• 57 files, +76 / −1450.

Problem

We do not need the benchmark runner UI for now, but it kept a footprint across the tab store, chat-panel surface state, sidebar connectors, session metadata, i18n and E2E config. On develop it was already effectively unreachable: no menu offered the benchmark create target, nothing created a benchmark tab except E2E seed helpers, BenchmarkTabSidebar had no consumer — the only live entry was clicking a legacy "Benchmark run coordinator" session in the sidebar. Dead-but-wired UI still costs typecheck/lint/test time and confuses future refactors (the previous PR moved its hooks around for nothing).

Solution

Moved to .archive/src/… (self-contained): features/BenchmarkPanel/, modules/WorkStation/shared/SidebarModules/Benchmark/, TabContent/renderers/benchmark.tsx, engines/ChatPanel/panels/BenchmarkRunBuilder.tsx, store/benchmark/, api/tauri/benchmark/, app/root/e2e/helpers/benchmark.ts, and the two E2E specs (.archive/tests/e2e/specs/core/).

Edited in place to sever: store/workstation/tabs/{types,tabFactory,storage,index,factories/*} (tab type/factory/allow-list), TabContent/registry.ts, SidebarModules/index.ts, SortableTab (BookLock branch), CodeSidebarHeaderActions, store/ui/chatPanelAtom.ts + types/ui/chatPanel.ts + chatPanelSurfaceReducer.ts (create target / content mode / surface kind), ChatPanel/{index,ChatPanelContent,ChatPanelEmptyContent,hooks/useChatPanelContentState}, NavigationSidebar/connectors/{useWorkstationSidebarHandlers,useSessionMenuItems/*,sidebarConnector.pinnedAndRevealData}, util/session/sessionDisplayMetadata.ts, app/root/{E2EBootstrap,e2e/types}, tests/e2e/wdio.conf.mjs, i18n/locales/*/sessions.json.

Deliberately kept: src-tauri/src/benchmark/ (Rust commands still compile; now unreferenced from the frontend — candidate for a backend-only follow-up), the Housekeeper token benchmark (housekeeperTokenBenchmark, integrations:housekeeper.benchmark.* — a different feature), and the generic flask-conical / folder-benchmark icon entries.

Potential risks

• User-visible behavior change: legacy "Benchmark run coordinator" sessions and their child sessions now appear in the sidebar as ordinary sessions (previously children were hidden and the coordinator opened the run list). They can be deleted like any session. Any persisted benchmark WorkStation tab is dropped by the storage allow-list on load (same handling as earlier archived tab types).
• The Rust benchmark_* commands remain registered; nothing calls them from the frontend anymore. Harmless, but dead until removed.
• No migration, no schema change. i18n edit is pure key deletion (verified per-locale diff).
• Stacked on refactor(hooks): dissolve src/hooks/workStation into owning modules #821 — the diff view will include refactor(hooks): dissolve src/hooks/workStation into owning modules #821's commits until it merges.

Test plan

• pnpm typecheck — 0 errors (also gated per commit by the pre-commit hook)
• pnpm lint over src/ — 0 errors
• pnpm test — 1099 files / 8679 tests pass (one exhaustiveness count in workspaceState.test.ts updated 39 → 38 for the removed tab type; sessionSidebarRow.test.ts benchmark-icon case removed with the feature)
• node --check tests/e2e/wdio.conf.mjs after removing the docker fixture builder
• rg for the archived module paths / benchmark tab-type literal across src/** → 0 hits outside the Housekeeper token benchmark and the Rust-command comment in api/tauri/rpc/schemas/cli.ts
• Not exercised: manual click-through; the archived surfaces were already unreachable so there is no live surface to check beyond "sidebar still renders sessions"

Submit checklist

• The PR is focused and has a scoped Conventional Commits title, such as feat(scope): summary or fix(scope): summary.
• I ran the relevant checks, or explained why they were not run.
• No secrets, private config, generated output, or unrelated formatting changes are included.
• Docs, screenshots, and locale updates are included when the change needs them.