# UX Goal Analysis Completion Audit

Status: implementation is underway; full goal is not complete.

Original goal:

> Analyse the complete UX and all usage scenarios you can think of, every single part of the system has to be analysed and tested. The goal is that all is functional, working, logical, usable, seamless, intuitive, well structured, and designed. Done when judges give us 9/10 - 5 different personas. First analyse, define all to be corrected, then after approval we will code.

## Evidence Artifacts

- Approval brief: `docs/audits/2026-07-08-ux-approval-brief.md`
- Complete UX audit: `docs/audits/2026-07-08-complete-ux-usage-audit.md`
- Route/scenario manifest: `docs/audits/2026-07-08-ux-route-scenario-manifest.md`
- Route evidence T11 analysis: `docs/audits/2026-07-08-route-evidence-t11-analysis.md`
- State/failure scenario matrix: `docs/audits/2026-07-08-ux-state-failure-scenario-matrix.md`
- Focused T12 state/failure analysis: `docs/audits/2026-07-08-state-failure-t12-analysis.md`
- Focused Mobile Executive T2/T12 analysis: `docs/audits/2026-07-08-mobile-executive-t2-t12-analysis.md`
- Non-main surface scope: `docs/audits/2026-07-08-ux-non-main-surface-scope.md`
- Five-persona scorecards: `docs/audits/2026-07-08-five-persona-judge-scorecards.md`
- Five-persona judge runbook: `docs/audits/2026-07-08-five-persona-judge-runbook.md`
- Correction register: `docs/audits/2026-07-08-ux-correction-register.md`
- Web Guidelines line findings: `docs/audits/2026-07-08-web-guidelines-line-findings.md`
- Runtime accessibility T10 analysis: `docs/audits/2026-07-08-runtime-a11y-t10-analysis.md`
- Visual T5 classification: `docs/audits/2026-07-08-visual-t5-classification.md`
- First-run onboarding T1/T2/T12 analysis: `docs/audits/2026-07-08-first-run-onboarding-t1-t2-t12-analysis.md`
- Shell overlay T10/T12 analysis: `docs/audits/2026-07-08-shell-overlays-t10-t12-analysis.md`
- Source inventory consistency audit: `docs/audits/2026-07-08-source-inventory-consistency-audit.md`
- Browser Companion T19 analysis: `docs/audits/2026-07-08-browser-companion-t19-analysis.md`
- Desktop wrapper T14 analysis: `docs/audits/2026-07-08-desktop-wrapper-t14-analysis.md`
- Admin/CLI utility T15 analysis: `docs/audits/2026-07-08-admin-cli-utility-t15-analysis.md`
- AI-tool hook lifecycle T16 analysis: `docs/audits/2026-07-08-ai-tool-hook-t16-analysis.md`
- Developer/substrate T17 analysis: `docs/audits/2026-07-08-developer-substrate-t17-analysis.md`
- Ops/deployment/CI/judging T18 analysis: `docs/audits/2026-07-08-ops-deploy-ci-judging-t18-analysis.md`
- Public launch funnel T13 analysis: `docs/audits/2026-07-08-launch-funnel-t13-analysis.md`
- Phase 1 implementation plan: `docs/superpowers/plans/2026-07-08-ux-phase-1-corrections.md`

## Requirement Trace

| Goal requirement | Current evidence | Status | Remaining condition |
|---|---|---|---|
| Analyze complete UX | Main audit covers shell, Home, workspaces, chat, memory, agents, skills/connectors/MCP/marketplace/launcher, team/admin/billing/vault/backup/compliance, Settings, desktop/mobile, static UX guidelines, and rendered test evidence. | Analysis ready | Keep audit updated as fixes land. |
| Analyze all usage scenarios we can think of | Route/scenario manifest lists route-level journeys; state/failure matrix lists account, tier, disclosure, model, data volume, offline, destructive, responsive, accessibility, and performance state combinations; the focused T12 supplement turns those into persona-specific state-bundle fields and correction candidates; the focused mobile supplement adds rendered 390 x 844 evidence for Home, Settings, Profile, Memory, workspace chat, Command Center, and Workspace Switcher; the first-run supplement exercises clean-data onboarding without skip flags through desktop completion and mobile Profile. | Analysis ready | Add new scenarios only if implementation or judge runs reveal missing cases. |
| Every part of the system analyzed and tested | Route manifest maps every registered route, major overlay, and the embedded/retired app surfaces found in source inventory; the focused T11 supplement now records the AppShell route registry, direct route-reference counts, app route-table command evidence, component evidence for command-only routes, ad hoc smoke for the previous zero/thin route set, a current all-route built-preview smoke covering 33 desktop routes plus 11 mobile spot-checks, and zero/thin route correction candidates. Focused overlay evidence now covers Keyboard Shortcuts, Persona Switcher, Spawn Agent, Workspace Switcher, Notification Inbox, Command Center, Create Workspace, Upgrade Modal, and source-only Context Rail/Onboarding Tooltip/Trial Expired states. Non-main surface scope maps `apps/www`, Tauri wrapper, installer/update, release workflow, active sidecar-startup paths, admin web, CLI launchers, marketplace CLI, memory/MCP utilities, AI-tool hook lifecycle packages, Browser Companion extension, SDK/server/worker/WaggleDance, substrate/compiler workspaces, deployment/ops config, CI, benchmarks, and judging artifacts. Additional non-main command/rendered evidence now covers `apps/www` tests/build, host-corrected local production launch-site route/API smoke, fresh public-site rendered Browser evidence, desktop TypeScript/static tests, static Tauri config/update tests, sidecar resource preflight, service-level E2E, utility/admin Vitest slices, utility TypeScript/build checks, source and built help/runtime smokes, MCP protocol smoke, hive-mind CLI test-discovery and help-side-effect evidence, T16 shared/agent/server route-contract tests, Launcher/prompt/adapter tests, root-run hook/shim package tests, hook/shim package-local scripts, hook package typechecks, compiled hook-bin help smokes, 13/13 remaining workspace no-emit TypeScript, agent/core/optimizer/weaver package tests, root-run SDK/substrate/shared/worker/WaggleDance/compiler tests, server full-suite plus isolated benchmark evidence, YAML/Compose config evidence, ignored env-file tracking checks, benchmark harness typecheck, root-run benchmark tests, Browser Companion source inventory, direct sidecar save evidence, current toolbar/extraction probes, legacy-trust extension save evidence, focused secure token-bootstrap/background-save coverage, secure-default loaded-extension save evidence, and Memory frame/UI confirmation. | Partially tested | T11 must still codify rendered-route evidence and add deeper workflow/state owners; remaining unsampled T10/T12 overlay/state findings must be fixed or deferred; T13/T14/T15/T16/T17/T18/T19 must fix or defer launch/desktop/utility/hook/developer/ops/extension gates before final claim. |
| Functional and working | In-app Phase 1/T10/T11/T12 evidence is much stronger: core accountless lane, shortcut/navigation, visual tracked views, mobile Settings/onboarding, route ownership, and five-persona state bundles have passing evidence. Launch/non-main gates still block a full-system claim. | Not achieved | Finish or explicitly defer launch, packaged desktop, utility/admin, hook, developer/substrate, ops/deploy, and Browser Companion gates; then run judge scoring. |
| Logical and usable | The register maps remaining logical/usable blockers to T10/T12/T13-T19, and current five-persona bundles pass route, failure, console, network, page-error, and visible-overflow gates for the accountless/no-LLM sampled lane. | Not achieved | Expand or defer real authenticated Teams server, destructive/failure, packaged desktop/launch, hook, utility, developer, ops, and Browser Companion evidence. |
| Seamless, intuitive, well structured, designed | Current sampled in-app lane is cleaner, including 0 visible overflow in the five-persona bundle and fixed mobile workspace/chat strip overflow. Final readiness is still below a full 9/10 until judge scorecards and non-main gates are closed or deferred. | Not achieved | Clear remaining T10/T12 gaps, non-main gates, and final judge protocol. |
| Five different judge personas score 9/10 | Scorecard artifact defines the five judges, scoring dimensions, caps, and evidence requirements; state/failure matrix defines the state bundle each judge must exercise; `tests/five-persona-state-bundle-contract.test.ts` now enforces that every persona scorecard declares account, billing, disclosure, model, data, offline/error, viewport, non-main gate decisions, Mobile Executive Command Center proof, Marketplace unavailable proof, Agents slow-list proof, Agents large-list proof, Memory slow-list proof, Timeline/Event large-list proof, Files upload failure/success proof, billing checkout unavailable proof, Team active billing proof, Team settings unlocked proof, billing checkout success/cancel return proof, backup restore-failure proof, approval grant revoke-all proof, local model unavailable proof, and Cockpit degraded-health proof; `tests/e2e/five-persona-state-bundles.spec.ts` now writes rendered accountless/no-LLM route plus sampled failure/workflow/scale/slow-data and overlay evidence under `output/playwright/five-persona-state-bundles/`. | Not achieved | Expand or defer real authenticated Teams server, real deployed/provider checkout success/cancel, packaged desktop/launch, hook, utility, developer, ops, and browser-companion evidence, then run judge protocol; all five must score at least 9/10. |
| First analyze, define all to be corrected | Audit, correction register, approval brief, and Phase 1 plan define findings, tickets, phases, files, tests, judge impact, and closure evidence. | Satisfied for approval | Keep the register current as corrections land and new judge evidence exposes gaps. |
| Then after approval we will code | User approved implementation; Phase 1/T10/T11/T12 corrections and guard tests are landing against the correction register. | Execution underway | Finish remaining correction-register items, capture rendered five-persona evidence, and run judge protocol. |

## Current Blockers To Full Goal Completion

1. The in-app accountless lane is substantially improved, but a full-system claim remains blocked by launch, packaged desktop, utility/admin, hook, developer/substrate, ops/deploy, and Browser Companion gates unless explicitly deferred.
2. Final five-persona judge scoring has not happened.
3. Real authenticated Teams server, real deployed/provider checkout success/cancel, packaged desktop, launch, hook lifecycle, remaining broader slow-data/performance, and remaining destructive/failure states are not yet fully evidenced or deferred.
4. Route evidence is explicit and regression-owned for the route-existence layer; T11 has current all-route built-preview evidence and codified focused route coverage. Remaining route-adjacent gaps are deeper workflow/state owners, runtime accessibility, and persona-specific failure/authenticated lanes tracked under T10/T12/T16.
5. T12 is explicit and partially evidenced but not closed: a focused state slice passes 9 files / 79 tests; chat SSE drop/retry coverage exists by source inspection; and the five-persona state-bundle E2E now passes 5/5 with route evidence, sampled failure/workflow/scale/slow-data probes, overlay evidence, screenshots, console/network/page-error checks, visible element-bounds checks, and deferral fields bundled together per persona. Current sampled rendered probes include chat backend offline, Memory API unavailable, Memory slow-list handling, Memory large-list handling, Wiki export failure, Timeline/Event large-list handling, Launcher sidecar offline, Cockpit health degraded, Agents slow-list handling, Agents large-list handling, Marketplace unavailable, Marketplace large-catalog handling, Files upload failure, Files upload success, Files large-list handling, billing checkout unavailable, Team active billing state, Team settings unlocked state, billing checkout success return, billing checkout cancel return, backup create failure, backup restore failure, approval grant revoke-all, local model runtime unavailable, and mobile chat backend offline. Remaining T12 closure needs real authenticated Teams server, real deployed/provider checkout success/cancel, packaged desktop/launch, hook lifecycle, broader slow-data/performance states, and explicit deferrals where applicable.
6. T10 is now both static and runtime-evidenced for the sampled core route and overlay set: focused axe/runtime gates, Command Center description/mobile fit, Notification Inbox semantics/mobile fit, Create Workspace hierarchy, shell landmarks, Files scroll regions, and many named-control/form-metadata fixes are covered. T10 remains open for broader unsampled form metadata, remaining unsampled icon-only controls, modal focus-return evidence beyond the covered overlays, and first-run/account-state polish that is still tracked under T1/T12.
7. Launch funnel, desktop wrapper, utility/admin, AI-tool hook, developer API/background/substrate, ops/deployment/CI/benchmark/judging, and Browser Companion extension gates are now explicit and partially command-tested or source-inventoried; T13 currently has a launch-scoped P0 from canonical `waggle-os.ai` and `www.waggle-os.ai` DNS failing to resolve, missing signed installer artifacts, checkout/deploy/legal proof gaps despite fresh clean-console localhost evidence, and Vercel/DNS proof still open; T14 has static/native-source evidence plus release/tray source hardening and source/unit/typecheck proof that native update/service-watchdog events surface as toasts, but still lacks packaged interaction smoke and packaged proof of update/service UX; T15 has passing source/test/build lanes, focused marketplace built help/invalid-command plus package manifest/packed-file alignment and clean installed packed-CLI `npx` help/invalid-command recovery, launcher help/invalid-port/occupied-port recovery plus packed first-command, clean installed occupied-port startup-recovery proof, and clean installed long-running `/health` startup, `@waggle/cli` built/packed/local package-closure installed `npx` help plus installed local REPL startup/slash-command/exit and streamed chat/provider plumbing, memory MCP built read/write plus local package-closure installed read-only startup, hive-mind MCP built read/write plus local package-closure installed read-only startup, hive-mind CLI local package-closure installed `npx` help, successful memory/hive-mind MCP protocol smokes, admin-web shell/mobile/deep-link/label fixes, and a package-owned admin-web rendered gate covering all seven pages at desktop/mobile widths plus browser back/forward traversal, page-level keyboard traversal, desktop/mobile visual snapshots, malformed analytics response recovery, all-page initial API-failure recovery, rendered mutation/destructive-failure recovery for capability policy save, capability override create/remove, capability request decision, member invite, member role change, member removal, and team settings save, plus local bearer-auth wrong-token/valid-token behavior, but registry-only proof after internal package publication remains open; T16 has strong route/contract, hook package, hook/shim package-local scripts, typecheck, compiled-bin, packed-package `npx` lifecycle, focused no-reconnect observed-output coverage, hook stdout/stderr/Backup-Recovery/Check-failed/uninstall-cleanup/More-output/structured-failure/empty-output/launch-only detail coverage, focused registry-aware third-party adapter launch coverage, partial rendered Launcher Browser evidence, codified Playwright standard install changed-file/pointer/backup/recovery labels, all six hook-capable install/verify/uninstall rendered transitions, sidecar-offline Retry, long-stderr summarization, and non-built-in adapter launch-only/prompt proof, Browser smokes where real Verify renders `verify failed (exit 1)` with retry/uninstall/reinstall guidance instead of `HTTP 400`, mocked Codex Verify renders manual-approval rows without raw `[FAIL]`, and mocked Codex Uninstall renders restore/cleanup labels without `Install pointer`, plus gated real-tool OpenClaw launch/output/exit/process-clear proof, all-six gated isolated `/api/tools/hooks` install/verify/uninstall proof, Windows npm-shim launch fixes, Windows hook `npx` resolution fixes, Codex WindowsApps launchability/recovery fixes, Tailwind motion-token warning cleanup, `shape-selection.ts` dynamic/static import warning cleanup, and startup JS chunk cleanup below Vite's 500 kB threshold. T16 remains open for packaged desktop hook-status evidence and remaining warning hygiene; T17 has broad package evidence but remaining non-hook package-local test-script failures and a full-server-suite perf flake; T18 has config/benchmark evidence but secret-safe ops logging, Render target, CI/infra, local Docker-engine availability, benchmark package-local tests, and current judging artifacts remain open; T19 now has source evidence, direct sidecar save evidence, content-script extraction evidence, legacy-trust extension save evidence, focused secure token-bootstrap/background-save coverage, secure-default loaded-extension extraction/save/frame confirmation, popup keyboard/focus/Enter save proof, direct popup Save page click evidence, restricted-page disabled-state recovery, honest memory-destination copy, context-menu handler coverage, stable packaged-ID pairing proof, Memory search provenance consistency for imported captures, existing chat `auto_recall`/catch-up imported provenance, sticky accessible popup recovery, and rendered Memory UI confirmation after secure popup saves, but still lacks native toolbar-bubble proof, native context-menu click proof or deferral, and signed Web Store/installer-distributed extension proof if release packaging is scored. Future recall result shapes need separate proof if scored. None of T13/T14/T15/T16/T17/T18/T19 is fully fixed or deferred.
   2026-07-09 T16 update: Codex WindowsApps launchability/recovery is now fixed at the UX contract layer. The restricted WindowsApps alias is detected as installed but not launchable, and Launcher shows recovery copy instead of a Launch button.
   2026-07-09 T16 route update: the isolated `/api/tools/hooks` lifecycle smoke now covers all six hook-capable tools (`claude-code`, `codex`, `codex-desktop`, `cursor`, `hermes`, `openclaw`) for install, verify, and uninstall.
   2026-07-09 T16 rendered update: `tests/e2e/launcher-rendered-states.spec.ts` now covers all six hook-capable tools rendering install, verify, uninstall, and refreshed hooks-active transitions. T16 remains open for packaged desktop hook-status evidence and warning hygiene.
   2026-07-09 T16 warning update: Tailwind ambiguous motion-token warnings are removed through named `duration-mo-*` / `ease-mo` utilities and `motion-class-hygiene.test.ts`; the `shape-selection.ts` dynamic/static import warning is removed through a static adapter import guarded by `build-warning-hygiene.test.ts`; the Vite large-chunk warning is removed by lazy-loading routes, closed shell overlays, ChatHost, and PostHog analytics, leaving the startup JS chunk at 421.96 kB minified / 114.08 kB gzip. Remaining warning hygiene covers color-env, embedding, and hook negative-path noise.
   2026-07-09 T14 event update: `TauriDesktopEventBridge` now mounts `listenDesktopShellEvents()` so `waggle://update-available`, `waggle://service-status`, and `waggle://service-restart-needed` become visible toasts; `tauri-bindings.test.ts`, `tauri-config.test.ts`, and `npm run typecheck:web` pass for this source/unit layer. Packaged watchdog/update smoke remains open.
   2026-07-09 T14 packaged-startup update: the debug Tauri binary now builds with `npx tauri build --debug --no-bundle`, launches from `app/src-tauri/target/debug/waggle.exe`, starts bundled `target/debug/resources/service.js`, and reaches `/health` with `status: ok`, healthy DB, and running watchdog. The smoke fixed invalid `plugins.dialog`, disabled updater runtime registration while updater config is absent, corrected bundled sidecar path selection, aligned bundled Node/native ABI, and allowed the Tauri `http://tauri.localhost` origin. MSI/WiX packaging, packaged tray interactions, close-to-tray, global shortcut, forced watchdog toast proof, installer trust, and signed-update proof remain open.
8. Five-persona judge run has not happened.
9. Judge scores are currently capped by known blockers.

## Completion Criteria Before Marking Goal Complete

The goal can only be marked complete after:

1. Phase 1 is approved, implemented, and verified.
2. Remaining P1/P2 work is either fixed or explicitly deferred without affecting judged journeys.
3. Route/scenario manifest has evidence owners for every registered route and major overlay.
4. Verification commands pass with current source and fresh-port browser evidence.
5. Desktop/mobile screenshots prove no critical clipping, overlap, or unusable navigation.
6. State/failure bundles are declared for every judge persona and each bundle has evidence or an approved deferral.
7. Launch funnel, desktop wrapper, utility, hook, developer/substrate, ops/deployment/judging, and Browser Companion extension gates are fixed and verified, or the user explicitly defers them from the five-persona score.
8. Five scorecards are filled with evidence and every persona scores at least 9/10.
9. Correction register shows no open P0 and no unapproved judge-blocking P1.

## Current Recommendation

Do not mark the goal complete. Continue focused implementation and evidence capture against the correction register, with the next meaningful gates being broader T12 failure/authenticated state coverage and the non-main T13-T19 gates unless the user explicitly defers them from the five-persona score.