Skip to content

backlog.md

Latest commit

 

History

History
280 lines (249 loc) · 50.1 KB

backlog.md

File metadata and controls

280 lines (249 loc) · 50.1 KB
Raw

Adaptive Learner Backlog

Daily-planning view of items outside the phase plan. The authoritative roadmap lives in ROADMAP.md; use this file for granular items + status.

Strategic Expansion: Content-Repository MVP

EXP-023 Phases A + B + C slice SHIPPED (v1.67.0; #118 / #122 / #124). Phase A: one user content repository (connect / validate / sync / cache / browse, both modes). Phase B: multiple repos (list / add / remove / reorder = precedence, legacy single-repo migrated), share via /add-repo deep link + QR, automatic technical validation (Trust 0→1, re-checked each sync), per-repo source filter, and private/coach repos via a per-repo token. Official repo stays the default + non-removable. Phase C slice: a curated recommended-repos.json drives a discovery section + one-click add

  • an "Officially recommended" (Trust 3) badge, plus local-only per-repo star ratings (the discovery fetch is gated until the catalogue is published — see commit 70108061). EXP-025 (author-provided lesson sets): AUTH-01 (book-metadata schema + validator, #529) + AUTH-02 (book-companion rendering, #531) SHIPPED. EXP-026 (fold user lessons into the content tree): UGC-01..07 SHIPPED (#97). Still deferred (need a shared backend): community ratings, Trust 2 (community-verified), central index server, coach progress aggregation, one-time invite tokens, Share-Wizard direct-push, and AUTH-03+ (cross-repo author publishing flow). See EXP-023.

See docs/explorations/ for the full strategic plan:

This expansion transforms Adaptive Learner from an AI-chat-only tool into a full learning platform with downloadable content sets, interactive exercises, and dual-mode operation (Content-only + AI-augmented).

State: post v2.16.0 (A reliability release. Errors that used to vanish now say what went wrong: a failed tutor reply shows a message with a retry instead of an empty bubble, a session whose history could not load says so instead of looking deleted, and a page that crashes while rendering offers a way back and a report button. Settings actions, set exports, the share check and lesson progress no longer fail silently. When the backend cannot be reached, you now read "No connection to the server" instead of a browser message. see changelog/releases/v2.16.0.md); prior post v2.15.0 (Exercises get deeper. A lesson can now declare variables whose values are drawn fresh for every attempt, with numeric free-text answers graded within a tolerance. Three new exercise types join the catalog: Hotspot, Parsons and Ordering, all authorable in the lesson creator. Exercises can also carry an explanation that appears after you answer. In the lesson creator you can write explanations by hand, or have the AI write them when you create a lesson from text. You can also mark the grammatical role of words so cloze exercises pick better blanks. In the browser version, you can apply the AI check's suggestions to your own sets. see changelog/releases/v2.15.0.md); prior post v2.14.0 (Sessions get a game layer: an optional Spielmodus adds combo streaks, flying points, checkpoints, answer physics, playful hearts/countdown tension, and its own sound set - stackable with two new arcade minigames (Lern-Memory + Snake, TicTacToe against a beatable heuristic AI, a Simon color-sequence game) unlocked through XP, plus flash rounds at set completion. The Lernfunke mascot gets unlockable color variants, and avatars get a preset gallery, frames tied to level/badge unlocks, and photo staging with a confirmation dialog on figure change. see changelog/releases/v2.14.0.md); prior post v2.13.0 (feature release 2026-08-25: EXP-050 exercise-type conversion complete with carried review history + AI-suggest; "Edit as a copy" / fork derivation / SRS-preserving re-import; correction round + screen-wake fixes; EXP-044-Option-C UI consolidation; visual pipeline trustworthy + deterministic (#2712-#2722); see changelog/releases/v2.13.0.md); prior post v2.11.0 (feature release: stable exercise identities + one-time local progress re-key migration (content corrections no longer orphan review cards), matching-exercise UI rework; first enforce_admins back-merge PR + registry-free per-arch image archives in the publish chain; see changelog/releases/v2.11.0.md); prior post v2.10.0 (security release: app binds to 127.0.0.1 on both delivery paths (launcher engine 0.26.0 + compose), opt-in LAN via ADAPTIVE_LEARNER_BIND_ADDRESS; see changelog/releases/v2.10.0.md); prior post v2.9.0 (launcher quits without a system tray (runtime tray fact), lesson order drives the learning sequence, per-lesson editing; see changelog/releases/v2.9.0.md); prior post v2.8.0/v2.8.2 (distribution switch: launcher pulls the published GHCR image instead of building, review-progress recovery #2161; v2.8.2 security patch: API-CSP white page #2197 + no debug default in the bare container #2198; see changelog/releases/v2.8.2.md); prior post v2.7.0 (tag + notes without artifacts, publish gate refused fail-closed #2178, superseded by v2.8.0); prior post v2.6.1 (patch release: desktop launcher engine docker-app-launcher 0.15.0 (Docker-permission self-repair #2024) then 0.16.0 with the localized launcher UI now in the binary #2031/#2032, the frozen-binary launcher.json + window-icon + branding fix #2027/#2028, the Create-Lesson book-template file-upload i18n fix #2018, a canonical feature-overview page #2019 + the #1918 mirror-locale port #2033; no code/schema change; see changelog/releases/v2.6.1.md); prior post v2.6.0 (feature release: session chat rebuilt on assistant-ui + cutover #1126/#2012, book-text file upload with chapter picker + multi-select sections + exclusion heuristic + batch generation #1927/#1950/#1953, lesson picker for multi-lesson sets #1979, dictation audio upload #1911, set visibility via engine manifest flag #1993 (engine -> 0.14.0), launcher context-aware Docker detection 0.14.1 #1966/#1977, CI hardening (required checks #1969, testid gate #1981, baseline auto-sync #1980, Docker smoke #1995), versionless help #1975 + em-dash sweep #2013; no app-side schema/API/data-model change; see changelog/releases/v2.6.0.md); prior post v2.5.0 (feature release: Create-Lesson as a full exercise authoring tool (editable core types #1845, manual add + multiple_choice #1853/#1888, discoverable generation #1848, extension-authoring wizard #1859/#1865, prompt migration on edit #1861), ext:al-dictation as the fifth extension type #1881/#1887/#1895, PWA-update + AI-key-vault consumed as npm packages #1873/#1868, lib/exercises consolidation #1862/#1867/#1877, PR-PFLICHT #1846 + TESTPLAN-PFLICHT #1858, iOS/WebKit hardening #1839/#1832/#1843, engine -> 0.13.1; no schema/API/data-model change; see changelog/releases/v2.5.0.md); prior post v2.4.0 (feature release: Create-Lesson authoring upgrade (textbook-text knowledge lessons #1745, edit + combine own lessons #1752, card image upload #1764, typeable exercise count #1761), free-text multiple accepted answers + AI second opinion #1803, AI key-import UX #1769, engine re-pin -> 0.13.0 / schema 1.8 with uploaded picture_choice images #1774, frontend god-file burn-down completed #1802, content-load robustness (unicode card ids #1809, friendly lesson-load error #1826, missing-profile guard #1788) + ghost-content lifecycle fixes #1818/#1825/#1822, i18n script-sanity lint #1759, Docker/Bun close-out #1772; app schema 1.7 -> 1.8 additive, no app-side schema/API/data-model change; release-gate hardening #1829/#1830/#1831; see changelog/releases/v2.4.0.md); prior post v2.3.0 (polish + consolidation: EXP-044 concern-split complete - global.css split byte-identically into per-concern styles/legacy/ files, now a pure import manifest (byte-identity gate #1657, final #1592 zone wrap #1695); lesson-player UX rework #1628/#1644/#1635/#1649; listen-first audio #1687; authored difficulty as adaptive cold-start prior #1683; hardened lesson/set file import/export #1681/#1685; Tile + FormHint primitives #1714/#1720; CI symmetric path gate #1659 + visual-baseline gate #1641 + forceRerunTriggers #1673; Create-Lesson + Dexie content-visibility fix sweep; no schema/API/data-model change; see changelog/releases/v2.3.0.md); prior post v2.2.0 (feature release: extension-exercise tier - 4 AI-authored types #1579 + native multiple_choice #1525, app as schema consumer of learn-content-engine (additive schema 1.5 -> 1.7), federated content-repo registry #1511, simpler mobile nav #1512, EXP-044 CSS-cascade repair (Tranches 0-2c), Playwright 1.61.1, Bun package manager, 499 lessons / 35 sets / 6 domains; see changelog/releases/v2.2.0.md); prior post v2.1.0 (polish - ghost-progress fix + Settings per-tab panels) + v2.0.0 (first public launch); earlier post v1.99.0 (mobile-hardening + content-flow: tappable multiple-choice ChoiceButtonGroup fixing the iOS tap-misfire #1341/#1342, opt-in auto-advance #1330/#1331, inline worked examples with additive lesson-schema 1.4 -> 1.5 #1326/#1327, "Ask AI" on theory/exercise via BYOK #1321/#1322, Discover source-language filter #1343/#1344 + UI-language follow #1347/#1348 + "New" indicator #1339/#1340, "Meine Inhalte" multi-select + bulk actions #1351/#1352, typed repo category + shared badge #1319/#1320, Latest-strand QR + QR image upload #1316-#1318, Soft Pop default theme #1315, Content-Engine source->canonical boundary EXP-042 #1311, device fixes - iOS focus-zoom #1353/#1354 + iPhone delete menu #1349/#1350 + persisted UI language #1333/#1334 + catalogue force-refresh #1337/#1338, i18n LLM quality pass id/ja/ko/tr/el #1313/#1314; no API/data-model change; see changelog/releases/v1.99.0.md); prior post v1.98.0 (ContentHub-redesign + i18n: collapsible desktop sidebar #1263, global grid/list content view default-list #1261 + Discover grid/list toggle #1264, per-set status management active/deferred/completed + delete in "Meine Inhalte" #1301, ContentHub inline info button #1274, "Weitermachen" panel removed from "Meine Inhalte" #1270, Dev Mode default ON in Latest / OFF in Haupt #1271, Error-Replay successes merge back into the SRS error list #1304, Spanish/French LLM-assisted i18n quality pass — es accents #1289 + fr informal "tu" register #1291 + 67 fr/26 es diacritics #1303, help+README/docs sweep #1280/#1283; no schema/API/data-model change; see changelog/releases/v1.98.0.md); prior post v1.97.1 (content-browsing UX + release-tooling: ContentHub IA reorg -> import/creation actions + My Lessons into the Import tab #1256, content-browser list/grid toggle #1243 + newest-download-first ordering #1242, Discover Search/Filter bar #1247, ContentHub info button #1252, key-export passphrase inline validation #1245 + key-vault password-manager opt-out #1239, one-click workflow_dispatch release actions #1233/#1234/#1235 hardened to run Playwright in the official container #1248/#1250; no schema/API/data-model change; see changelog/releases/v1.97.1.md); prior post v1.97.0 (schema + tooling release: lesson JSON-Schema as the App-authoritative single source of truth EXP-039 — generated schema/quality-rules/doc/TS-types from the Pydantic models #1197, TS lesson types consume the generated schema #1200, ajv-hardened structural validator #1205, make sync-schema-check gate; cloze multiselect #1195; exam-mode SRS interval boost #1040; passphrase-encrypted .alk key/provider export #1165/#1181/#1183; vertical desktop sidebar #891; single-set deep link /content/set/:setId #892; native Web Share sheet + desktop fallback #1219; Latest-strand visibility #1172; iOS Add-to-Home-Screen guidance #1167; EXP-041 exercise-type-vs-learning-goal generation rule #1225; device fixes #1216/#1218/#1227/#1228/#1211; #809 splits #1206/#1210, lib/ai split #917/#1190, verify-theme gate #1169, docs sweep #1215; no schema/API/data-model change; see changelog/releases/v1.97.0.md); prior post v1.96.0 (imported-chat backend/API-mode parity #1154 on top of the Dexie-mode imported-chat fixes #1078/#1122/#1137/#1141/#1143/#1147, Reverse #1013 + Endless #1015 lesson modes, invitation-code content sharing #1094, online-to-local data migration #1099, the docker-app-launcher engine #1064, SEO #1108, an auto-deploying preview site on adaptive-learner-content-test #1135; see changelog/releases/v1.96.0.md); prior post v1.95.0 (learning-modes release: Practice/Exam/Timed/Zufall lesson modes + "Fehler trainieren" entries #1007/#1008/#1010/#1026/#1025, exam delayed feedback + dedicated result view + 1.5x XP with persisted lesson_mode (Alembic 0033) #1037/#1039/#1033, word-tiles + free-text answer/solution toggles #1006/#1020, GitHub-repo set export #1018, launcher overhaul — single persistent window #1045 + verbose startup cleanup #1052 + make launcher-test #1051; see changelog/releases/v1.95.0.md); prior post v1.94.1 (launcher hotfix: Windows port check sets SO_EXCLUSIVEADDRUSE so an in-use port is no longer reported free + the Windows launcher .exe attaches #990; see changelog/releases/v1.94.1.md); prior post v1.94.0 (launcher-hardening release: Actions layer + CLI/GUI parity #966, a single never-self-closing launcher window + inline live-validated port field #984, end-to-end launcher fixes — prod build npm ci #954 / startup crash #961 / uninstall-verify #964 / repo slug + links #950/#952 / default port 8501 #962 / health-check on the published port #985 / local-repo build from a source checkout #981; lesson retry-with-improvement #984; Matching answer/solution toggle #978; no schema/API/data change; see changelog/releases/v1.94.0.md); prior post v1.93.0 (launcher + content-dialog release: Docker-first launcher flow + progress window + configurable port + brand icon #942, scrollable/dismissable AI/quality dialogs via ModalShell #937, AI content-check provider/model provenance #940; see changelog/releases/v1.93.0.md); prior post v1.92.0 (Dexie-hardening + technical-debt release: CRITICAL installed-PWA-uses-Dexie fix #907, browser-direct Dexie fixes #902/#903/#914, tabbed Dashboard #858, six god-folder splits #874–#879, cc>20 complexity offenders cleared #884–#887, ~387 docstrings #888, Block 9 cleanup #889; see changelog/releases/v1.92.0.md); prior post v1.91.0 (UX release: Navigation restructuring EXP-037 #850 — primary nav 12+→7 grouped entries via NavGroup + mobile BottomTabBar + ProgressHub + DiscoverHub Import tab, redirects keep old links alive; PWA update-banner dismissal now persisted #846; no schema/API/data change; see changelog/releases/v1.91.0.md); prior post v1.90.0 (feature release: AI Exercise Generation Pipeline EXP-036 / AIX-01..06 #817/#829/#834/#835/#836/#838; animated Matching pair resolution #825; per-provider Test button #814; desktop auto-update checker #843; AI session replies in the active UI language #828; localStorage snapshot in the .alb backup #791; localStorage→Dexie migration Part B #837; see changelog/releases/v1.90.0.md); prior post v1.89.0 (feature release: configured AI-providers overview in Settings with a masked key preview #812; .alb ZIP backup now carries a localStorage snapshot on export/import #792; browser-direct Anki AI card extraction in Dexie mode #808; theory-only offline lessons can be saved #796; changed — AI chat analysis runs in the active UI language #806, the API-key live test uses the models-list endpoint instead of a generation call #800, Gemini/OpenAI key-format validation relaxed #794/#782, the AI gets lesson-progress context on session start #798, directory restructuring grouping oversized directories by concern + a god-folder CI guard #811; fixes — native confirm() replaced with an app confirmation modal #784, button-styled set/learning-path links carry data-slot so labels stay visible #780, Dexie-mode api_key_backups no longer crash an API-mode import #788, .alb files accepted at every backup-import surface #786, per-level statistics labels stack above the bar #805, ContentPage complexity gate restored #790; deps — cryptography 49 + backend/frontend minor-patch groups #776/#777/#778; no schema/API/data change; see changelog/releases/v1.89.0.md); prior post v1.88.0 (feature release: QR-code app sharing — "Share the app" entry in the About tab shows a scannable QR code of the public app URL (copy / download-PNG / native share) via a reusable shared/QrCodeModal reusing the bundled qrcode lib #775; Content Browser / Discover search separation — the Content Browser shows only locally downloaded content, index search + per-set download moved to /discover, cross-linked both ways #773; media priority + companion book shown first in "Vertiefe das Thema" #770; fix — Settings token/API-key inputs no longer trigger the browser password manager #768; docs — release-time version-badge + roadmap/backlog header refresh #771; no schema/API/data change; see changelog/releases/v1.88.0.md); prior post v1.87.0 (feature release: EXP-034 content discovery + search index (#735/#738/#740/#742), Curriculum Builder (#737), daily learning reminders (#744), EXP-031 .alb ZIP backup (#714), EXP-032 deterministic content-quality checks (#712), Korean + Indonesian UI + social sharing + Anki .txt export (#728), configurable review-session length (#724), editor right-click context menu (#753), new teal brand mark + dark variant (#719/#725), Hindi UI complete (#755/#758); late hardening — mobile nav 375px overflow (#759), dexie backup .alb assertion (#760), live review-badge decrement (#761), equal-height choice tiles (#762); no schema/API/data change; see changelog/releases/v1.87.0.md); prior post v1.86.0 (feature release: EXP-033 AI Content Validation — set-wide AI checks + signature + AI-checked badge #676/#683/#687/#691; EXP-029 Media Integration — media.yaml + reciprocity gate + lesson media section + set badges + YouTube thumbnails #688; Settings "Check for updates" button #669; autofocus in Cloze/Free-Text #693; P1 fixes — review-badge decrement #675, SRS review dedup #665, mobile hamburger close #667; repository-UI i18n in 9 langs #662; Greek i18n placeholder #702; #689 CI-gate split — content-loader-dexie-ai extract + ContentPage cc baseline #704; docs — Library-First audit #695, Content-Repo Guide #670, 32-EXP status audit #677, roadmap/backlog restructure #681, EXP-032/033 design docs #658/#671; no schema/API/data change; see changelog/releases/v1.86.0.md); prior post v1.85.0 (P0 hotfix: PWA update banner rebuilt — bottom-anchored + AA-pinned --bg-surface/--fg-primary pair, env(safe-area-inset-bottom) padding + z-[9999], accent update button, 44px targets, regression-pinned #653; complexity ratchet gate unblocked by baselining the pre-existing FreeText/WordTiles cc-21 offenders #656; no schema/API/data change; see changelog/releases/v1.85.0.md); prior post v1.84.0 (maintenance: P1 content-repo import CORS-safe + retry + progress + dialog a11y #645; P2 inline Markdown in exercise prompts + labels #647; P2 clickable Settings profile picture with preview dialog #639; P2 backup import declines foreign JSON files gracefully #643/#641; user-repo import E2E test #637; EXP-031 ZIP-backup-format design exploration #644; no schema/API/data change; see changelog/releases/v1.84.0.md); prior post v1.83.0 (maintenance + gap-hardening: P1 review-session fix — Enter + element de-dup + live XP badge #631; P2 theory markdown-table styling #633; P2 theory back-link by topic #635; gap-hardening audit #630 — cloze Tab nav #623, hints feature-not-available policy #624, Dashboard favorites Top-5 cap #625, review "repeat in 2 days" #626, free-text "Fast! Achte auf:" typo hint #627, Dashboard Quick-Review button #628; no schema/API/data change; see changelog/releases/v1.83.0.md); prior post v1.82.0 (hint economy with XP cost + SRS impact + statistics #611 (Alembic 0030); smart review queue — weakness tiers + attempt history + cross-lesson mix capped at 20 #612 (Alembic 0031); PWA update prompt — polled version.json + SW prompt mode #614; Test Impact Analysis on PR CI — vitest --changed + pytest --testmon, full suite nightly + release #617; architecture-doc audit #620; recommended-repos discovery E2E reactivated #610; manual test plan automated — 52 Playwright specs across 7 sessions + nightly workflow #621; additive Alembic 0030+0031 on element_errors, Dexie schema unchanged; see changelog/releases/v1.82.0.md); prior post v1.81.0 (biggest feature release: Learning Statistics #584, SRS viz #592, staged hints #595, favorites #598, review UX #602, Dashboard gamification #583, lesson UX #589, offline UX #605, keyboard shortcuts #587, Hindi UI #571, LanguagePicker #568, event recording EXP-028 #566, avatar crop #560, Settings sidebar #549, editable username #579, gamification API #573, recommended-repos live #574, 20+ shared/ components; CI night-shift #552/#576; starlette CVE #607; EXP-027..030 docs; see changelog/releases/v1.81.0.md); prior post v1.80.0 (EXP-026 user-lesson folding into the content tree #97 + EXP-025 book companion #142 (AUTH-01/02) + user profile picture #508 + localized split titles #512 + a11y #514/#515 + maintenance #511/#537/#540/#541; see changelog/releases/v1.80.0.md); prior post v1.79.0 (minor: XP visibility #505/#510 + bidirectional matching selection #507/#509 + complexity burn-down complete (final offender #497, baseline empty) + radon hard gate Phase 2 #494/#495 + plugin-tests CI job (1018-test suite) #471 + reusability policy + shared primitives #474/#477 + P1 matching duplicate-pair scoring fix #480/#481; no schema/API/data change; see changelog/releases/v1.79.0.md); prior post v1.78.0 (maintenance/code-hygiene: complexity burn-down under the Phase 2 hard ratchet #408; the 1156-line session routes.py split #412 (last backend cohesion-baseline entry; .filesize-baseline now empty); a behaviour-preserving burn-down batch #416-#460 (LessonPage 67->12, ImportDetail 58->18, backup-diff previewRow 54->3, buildContentSetRow 49->13, ApiKeyRow 45->9, MatchingExercise 45->17, build_analysis_context 33->3, formatEventLog 40->4, NextStepSuggestions 36->14, plus the CCW frontend batch); two governance rules - Release-Freeze #410 + No-Amend-on-open-PR #412/#414; two flaky-test fixes - lesson-tts #165/#425 + pytest-randomly reactivated #426/#429; EXP-025 (Refs #142) + EXP-026 (Refs #97) design docs; no schema/API/data change; see changelog/releases/v1.78.0.md); prior post v1.77.0 (architecture release: the three-phase Dexie read-modify-write data-integrity remediation #390 -> #395/#398/#402; completion of the god-file decomposition across backend + frontend #341/#353/#354; three warn-only CI watchers - cohesion #371, security-scan #378, complexity #405; the Vibe Coding Policy #383; backend parameter-dataclasses #376/#382; a feature-state policy #336; the gitflow branching model #334; an npm-audit fix #379; no schema/API/data change; see changelog/releases/v1.77.0.md); prior post v1.76.0 (maintenance: a useControlledExercise hook + a shared ExerciseFooter remove ~300 lines of lifecycle duplication across all 5 exercise renderers, zero behaviour/test change #322/#323-#328; the per-user backup/sync ownership check deduplicated into one canonical pair in sync_service (row_belongs_to_user + record_belongs_to_user) - security-directional drift closed, behaviour-preserving on all reachable data since every mapped user_id is non-nullable #329/#330; both backup buttons unified onto one saveBackupToDisk helper with a cross-component parity test #331/#332; TTS no longer reads raw Markdown aloud #320/#321; the Dexie-mode CI gate runs in the mcr.microsoft.com/playwright container instead of the hanging cdn.playwright.dev download #317/#319; the misconfigured prettier-frontend hook removed #316/#318; content-set download accepts the content repo's optional domain_label field, fixing a 400 on the psychology set #333; no schema/API/data change; see changelog/releases/v1.76.0.md); prior post v1.75.0 (minor - TipTap editor stack migrated v2 -> v3: the whole @tiptap/ stack (23 packages) bumped 2.27.2 -> 3.26.1 in one atomic change #311/#314 (a mixed v2/v3 tree does not compile - the failure mode that defeated the prior piecemeal Dependabot bumps and prompted the major-hold); only the v3 breaking-API deltas as code changes, verified against the installed 3.26.1 packages - TextStyle/Table default -> named imports, StarterKit's bundled Link/Underline disabled, setContent({emitUpdate: false}), generic NodeViewContent<"code">; 3953 vitest + 88 dexie-smoke green, zero test changes; Dependabot @tiptap major-hold removed #305/#306; pre-existing @eslint/js-10 no-useless-assignment break fixed in 5 files #312/#313; the two migrated editor files restored to the repo 4-space style after an inadvertent prettier reformat #315; no schema/API/data change; see changelog/releases/v1.75.0.md); prior post v1.74.0 (maintenance/infrastructure: 438 ESLint warnings cleared + gate tightened to --max-warnings 0 so Frontend Tests CI is green again #220/#292; dexie-smoke E2E CI workflow #301/#302; @tiptap/ reconciled to a consistent v2 tree (2.27.2) #304 + Dependabot @tiptap major-hold #306; CI installs with --legacy-peer-deps #293/#294; Dependabot bumps - feature-strategy 0.1.1 / @types/node 25 / @vitejs/plugin-react 6 / @eslint/js 10 / ruff / minor-patch groups #295-#300, #307-#310; visual baselines unchanged; see changelog/releases/v1.74.0.md); prior post v1.73.0 (minor - feature-strategy integration: central registry + strategy replaces ad-hoc API-key gating with active/disabled/hidden states + a reactive {mode, hasAiKey} context (#287); E2E feature tests - content-repo import #278, Anki #279, NotebookLM #280; NotebookLM AI buttons key-gated #281; TipTap extensions pinned back to 2.27.2 to fix the v3/core-v2 MISSING_EXPORT build break #267/#288; Progress.test wrapped in TestFeatureProvider #289/#290; dependency bumps #258-#266; see changelog/releases/v1.73.0.md); prior post v1.72.2 (patch — Import-Detail + Import inline styles -> token-backed Tailwind #275; Anki empty state with icon/title/body/import-CTA/API-key-notice in 8 langs #276; all 3 frontend import cycles resolved via pure type modules (content-validation-types / storage/export-types / api/request-types), madge 0, check-circular baseline 3->0 #252; see changelog/releases/v1.72.2.md); prior post v1.72.1 (patch — dark-theme button-background fix #271 (raw no longer inherits UA buttonface ~#efefef on dark themes; base-layer background-color:transparent, the unfinished half of v1.71.0's #185) + 5 serious axe a11y fixes #273 (aria-progressbar-name on mission/XP bars, nested-interactive on the ProfileRadar role=img wrapper, listitem on Content knowledge groups, color-contrast via #271 — 0 violations on all 7 audited routes) + visual lesson-matching baselined via lesson-check #270 + the axe suite now runs via browser.newContext() #272; see changelog/releases/v1.72.1.md); prior post v1.72.0 (minor — CI/quality-infrastructure build-out: ESLint (#222), security scanning (#224), coverage floors (#225), madge circular-dep guard (#251), Dependabot (#253), Prettier step 1 (#254), bundle analyzer (#255), Stylelint (#257), Playwright visual regression (#244) + axe-core (#246); plus dark-theme contrast/layout + Import-Detail UX fixes (#226/#228/#230/#232/#234/#236/#238/#240/#242); see changelog/releases/v1.72.0.md); prior post v1.71.1 (patch: dark-theme contrast & spacing sweep + Session-Detail export fix #209 + bare .btn base colour #211; see changelog/releases/v1.71.1.md); prior post v1.71.0 (minor: clearer Matching result feedback - wrong pair shows "Deine Antwort" (red/X) + "Richtige Antwort" (green/bold/check) on separate lines, correct pair confirms "A -> B" with a green check, new your_answer + reworded correct_hint i18n in 8 langs via AA-pinned matching tokens (#191); Enter works in the lesson-end correction round (CorrectionBlock) via a controlled cloze + shared useLessonEnterKey hook (#187); systematic dark-theme button contrast - base-layer button{color:inherit} so raw buttons stop falling back to UA black with preflight off, losing to explicit text-* utilities (#185); content-repo token password input wrapped in a form (#119); test isolation - content-loader cache reset between tests fixes the flaky cross-identity backup round-trip under pytest-randomly (#164) + lesson-tts Dexie-gate spec headroom/retry (#165); content fix - Miller "7 ± 2" cloze accepts keyboard-typeable +/- forms (content #33); see changelog/releases/v1.71.0.md); prior post v1.70.2 (patch: theme contrast + Matching-exercise fixes - soft-pop secondary button white-on-teal fixed + every button-variant colour pair pinned across 12 themes (#179); red removed from Matching pair colours via a dedicated red-free --matching-pair-1..7 palette (#181); Matching result state distinguishes correct/wrong pairs - badges persist, correct green both tiles, wrong red both tiles + correct-partner hint, unmatched neutral, new per-theme --matching-correct/-error tokens AA-verified (#183); see changelog/releases/v1.70.2.md); prior post v1.70.1 (patch: onboarding / assessment / content-browser / landing UX fixes - uniform onboarding-wizard step height (#169); assessment first-step "Continue later" exit + replace-navigation history fix so browser-back doesn't land on the onboarding form (#171); landing docs link opens in a new tab (#173); onboarding Name -> Topic tab order, Topic help icon out of the tab order (#175); Content Browser secondary buttons ghost -> outline so they stay visible at WCAG AA in dark themes (#177); see changelog/releases/v1.70.1.md); prior post v1.70.0 (minor — first-run backup restore on an empty install (#150) + a comprehensive v1.61-v1.69 MkDocs documentation overhaul in all 8 languages, deployed at https://astrapi69.github.io/adaptive-learner/docs/ (#157) + context-sensitive in-app help (the nav "?" opens the current view's entry, help articles gain a "Learn more" link to the docs site, 9 new glossary terms in all 8 languages) (#159) + a manual pre-release QA test plan at docs/reference/MANUAL-TESTPLAN.md; see changelog/releases/v1.70.0.md); prior post v1.69.0 (minor — theory example links (#139) + per-domain book recommendations (#141) via Sprint 2 (#153); Enter-key shortcut in the Error-Replay runner via a shared useLessonEnterKey hook (#154); backup-restore title fix — restore parses the manifest sets[].title so a restored set keeps its title (not the analysis- set_id) + step progress (#134), proven by a real export->import round-trip; see changelog/releases/v1.69.0.md); prior post v1.68.0 (minor — lesson-result export + theory back-links (#138/#140); matched-pair color + number-label overhaul (#145); dark-mode contrast fixes — button-styled anchors keep their variant color (#146) + outline/ghost buttons set an explicit foreground (#148); read-aloud no longer reflows the theory panel (#147); domain-aware matching wording for knowledge lessons (#149); search-icon-right + uniform exercise-card heights (#143); About AI-assistance credit naming Claude (Anthropic); see changelog/releases/v1.68.0.md); prior post v1.67.1 (patch — restored user-generated lessons keep their title + step progress: Dexie-origin user sets store the title in the contentSets row with no manifest.yaml, so the API restore landed them manifest-less (Dashboard showed the set_id, step count collapsed to "Fortsetzen"); restore now synthesises a one-set manifest from the backup metadata (#134); + EXP-024 Phase 1 backend repository-pattern refactor, internal only (#133); see changelog/releases/v1.67.1.md); prior post v1.67.0 (backup is a complete self-contained snapshot + EXP-023 user content repositories: full 30-table export (empty=[]) + restore logging + scrollable per-table import summary + persistent error toasts (#126); subjects reconcile by (parent_id,name) (UNIQUE + Alembic 0028 dedup, null-aware root match, self-ref remap) — no fresh-install taxonomy duplication (#127); cross-identity re-home so a prior-identity backup lands, 0 errors vs 192 (#129); downloaded content sets travel in the backup (wire 1.3.0, both modes) so restored lessons open + user-generated sets aren't lost (#130); EXP-023 multi user content repos A/B/C (#118/#122/#124); Chrome password-form console cleanup (#119); new BACKUP-AKZEPTANZTEST gate; see changelog/releases/v1.67.0.md); prior post v1.66.0 (maintenance — systematic backup-restore fix (generic unique-key matching for 13 non-id-UNIQUE tables + FK-graph child remap + placeholder reclaim; JSON serialization for dict/list values in Text columns, fixes type 'dict' is not supported; empty-table skip on export+import; verified end-to-end through the real backup API) (#115, #117); GH-Pages stale-deploy chunk-reload via lazyWithReload + cleanupOutdatedCaches (#113); subject-filter UX — hide at <=1, most-used-first, category groups above 5 (#111); see changelog/releases/v1.66.0.md); prior post v1.65.0 (resumable assessment — abandon-and-resume with project-scoped localStorage progress + Dashboard/Settings Continue/Create/Retake invites, cleared on completion (#106); Enter-key lesson shortcut — check then advance, free-text/cloze submit on Enter, Settings>Learning toggle (#103); design-token architecture — global.css colors routed through tokens + no-hardcoded-colors guard extended to CSS + Tailwind palette + docs/policies/DESIGN-TOKENS.md (#101); matching exercises — blue/green side tints + A/B chip via color-mix() tokens, AA-verified (#108); fix — assessment result radar/badge overlap (#105); see changelog/releases/v1.65.0.md); prior post v1.64.0 (onboarding overhaul — a two-field quick start (name + topic, the rest defaulted) + an optional one-question-per-screen profile wizard (OnboardingWizard: goal / timeframe / daily minutes / current problem / opt-in assessment), assessment now OPT-IN ("Jump right in" -> Dashboard; assessment only from the wizard's last step) (#92, #94); fixes — Content Browser single scrollbar (html/body lock both axes so #root is the sole scroll container) (#42) + sticky lesson footer pinned across steps (lesson-page fills the viewport, step grows flex-auto; regression pin e2e/dexie/lesson-footer-stability.spec.ts) (#43) + a WCAG contrast pin for --accent-as-text + a catppuccin-mocha nudge so all 12 themes pass computationally (#96); see changelog/releases/v1.64.0.md); prior post v1.63.0 (6 recommended WCAG-AA theme presets + systematic i18n audit (#80 — subjects.* data-i18n for 77 seeded names + 92 missing t() keys incl. the whole editor.* toolbar, all 8 langs) + dashboard subject filter scoped to the user (#72) + theme/i18n fixes (#82/#84/#87); see changelog/releases/v1.63.0.md); prior post v1.62.0 (backup-restore data-integrity hardening (#57 datetime coercion + #64 orphan-FK-skip) + GitHub-Pages build provenance (#66) + content cache-bust (#62) + UI/i18n conformance (#51/#55/#76/#53/#68/#78/#69) + .claude governance + Bibliogon templates/labels; see changelog/releases/v1.62.0.md); prior post v1.61.0 (app-wide shadcn button conformance (~200 buttons across 13 page areas) + lesson resume-at-paused-step (LessonProgress.current_step, Alembic 0027 + Dexie) + cross-repo content validation (validate_bundled_content.py + a Content-stats CI gate + README CONTENT-STATS block, 330 lessons / 16 sets) + backup-restore fixes (badges.key natural-key upsert + user_badges.badge_id remap, FK-topological _RESTORE_ORDER, imported_conversation_id in backup columns); see changelog/releases/v1.61.0.md); prior v1.60.0 (lesson-reading UX + Learning Path Achievement Map + Tailwind exercise renderers + help-glossary perf + B1 content complete: auto-hide lesson header on scroll (useScrollDirection on #root, motion-safe Tailwind transform, sticky footer stays); Learning Path Map view (domain-grouped; 3 views: Persönlich / Map / Graph); Settings icon-only mobile buttons; all 5 exercise renderers migrated to Tailwind (~85-90 %) + theme audit (43-token parity, themed Dialog overlay) + dead-CSS removal; help glossary lazy per language (main index chunk 731->449 KB raw / 245->138 KB gzip, PERF-HELP-GLOSSARY-LAZY-01); same-language imports auto-detected as knowledge domain (Save stamps lesson domain, Share Wizard inherits the pair; E2E variant 2 un-fixme'd) + Dexie async-load wizard fix + github_service mypy cast; B1 content complete — de->es/en/fr B1 (15 each), 271 lessons / 13 sets / ~66 h); prior v1.59.0 (Learning Path Redesign — personal path with zoom levels: two-level view replaces the all-225-lessons xyflow graph as the default (Level 1 set rows sorted by last activity with progress dots + action; Level 2 accordion lesson detail with stars/mastery/dates + adaptive/retry-errors); "Nur meine" / "Alle Sets" filter + collapsible not-downloaded section, both persisted; next-CEFR-level offer on completion; xyflow kept as a lazy-loaded alternative view, removed from the default bundle; nav hamburger moved left on mobile); prior v1.58.0 (user-centric UX overhaul: Continue Learning ("Weitermachen") section on the Content Browser + Dashboard surfacing the most recently-touched lesson per set; Content Browser reordered search-first with icon-only mobile action buttons; Dashboard reordered Continue-Learning-first; responsive icon/text button pattern); prior v1.57.0 (community PR automation — fork -> commit -> PR for community sharing, GitHub PAT in Settings -> Integrations; Content Browser instant search; Tailwind Phase D; psychology to 90 lessons); prior v1.56.0 (performance + PWA hardening: ~460 KB gzip saved via lazy i18n catalogs + curated highlight.js, bundle audit, Dexie/backend queries verified healthy; offline indicator + network-aware buttons, background-sync queue, cache-management UI, install prompt, API-mode lesson caching; carries the Tailwind/shadcn migration Phases B-D + backend rate-limiting/OpenAPI; restored per-theme read-aloud in Dexie mode); prior v1.55.0 (Tailwind CSS v4 + shadcn/ui foundation (Phase A) + Error Replay; prior v1.54.0: import-time language pipeline + big content release; prior v1.53.0: content schema v1.3 + Python course + domain support — 7 sets / 100 lessons; code rendering + code-aware exercises; analysis-to-lesson source-language fix).** Prior: **v1.52.0 (DE->EN A1 content — 5 sets / 75 lessons —

  • BACKUP-API-RESTORE-01 backend restore-coverage fix, production DB guard, and the Lesson-Creator resumed-draft P0 fix).** Prior: **v1.51.0 (Phase 66 / EXP-022 — Visual Learning Path
  • Dexie backup overhaul: File System Access API save-to-disk, a "Your backup contains" preview, and a data-loss fix restoring the 10 gamification/progress/SRS/missions tables the Dexie export had silently dropped; BACKUP-API-RESTORE-01 filed P1). Prior: post v1.49.0 (Phase 65 — API-key UX + Community Sharing via PR + Analysis loading: API-key format validation + live Test button + rollback/restore cache (ApiKeyBackup, Alembic 0025, Dexie v24), stable secret.key Fernet source + UI-editable secrets.yaml keys; community sharing opens a GitHub pull request; chat-import Analyze loading indicator with real Cancel; friendly voice/mic errors. Plus, unreleased on main: native help translations in 6 languages — HELP-CONTENT-TRANSLATIONS-01).** Prior: v1.48.0 (Phase 64 — Community Sharing UX + Smart Lesson Organization: four-step share wizard, placement engine, duplicate/variation detection, author credit, contribution history, Missing-Lessons suggestions; content schema 1.2 -> 1.3; merges Smart Next-Step Suggestions after lesson completion + EXP-021 Lesson-Creator exploration). Earlier: v1.47.0 (Phase 63 — Lesson Flow Control: pause/abandon/resume + autosave + auto-resume, paused-lessons widget, lesson splitter; Word Tiles @dnd-kit reorder, mobile overflow fixes, backend CSP pass). Phase history through Phase 64 + per-release notes live in changelog/releases/. 30 tables on the sync surface. 13 plugins, 30 SQLAlchemy models, 1215 + 1018 + 4139 = 6372 tests green (backend + plugins + Vitest; verified 2026-06-15). Closed in this release line: BL-04 (QR scan, v1.7.0), BL-05/06 (sync gaps, v1.8.0), BL-07 (subjects/tags, v1.9.0), BL-08 (gestures, v1.10.0), BL-09 (model picker, v1.11.0), BL-10 (backup compare, v1.12.0), BL-11 (PT/TR/JA native, v1.13.0), BL-12 (TipTap, v1.14.0), BL-13 (E2E expansion, v1.15.0), BL-18 (gamification, v1.16.0), BL-21 (Anki, v1.17.0), BL-20 (voice, v1.18.0), BL-22 (NotebookLM, v1.19.0), BL-25/26/27/28 (import parser audit, v1.19.x), BL-29 (metadata.created_at ISO normalisation, v1.19.2), BL-03 (pluginforge-app-template, shipped externally), BL-30 (Git-Backed Learning Repository, v1.26.0). Phases 35-41 (v1.21.0..v1.25.0) shipped without consuming BL-IDs: Phase 35 doc-staleness refresh, Phase 36 import bugfixes, Phase 37 error-toast + GitHub-issue framework, Phase 38 in-app help system, Phase 39 WCAG 2.1 AA, Phase 40 release-automation hardening, Phase 41 identity persistence
  • Danger Zone.

Items ordered by impact and dependency chain. P0 = next up, P5 = speculative. Within each tier, smaller-scope and unblocking items come first; alphabetical-by-ID as final tiebreaker.

Audit-getriebene Priorisierung (aus journal/exploration-status-audit.md, 2026-06-17). Jetzt / Nächste Phase / Deferred führen die Task-IDs; die ausfuehrlichen Beschreibungen stehen in den jeweiligen Explorations-Dokumenten + in den P3-P5-Detailabschnitten unten. Kuratierte Sicht: ROADMAP.md.

Jetzt

Nächste Phase

  • EXP-025 — Content-Repo-Oekosystem: AUTH-03..AUTH-09 (AUTH-01/02 ausgeliefert #529/#531; Teil braucht geteiltes Backend → Deferred).
  • EXP-027 — weitere Sprachen: I18N-06 (Indonesisch), I18N-07 (Italienisch), I18N-08 (QA-Pipeline), I18N-09 (Content-Sprachpaare), I18N-10 (Exercise-RTL-Audit), I18N-12 (Arabisch-Content).
  • EXP-022 — Lernpfad-Ausbau: UC2 (Content-Browser als Graph), UC4 (Beziehungs-Editor).
  • EXP-021 — Lektions-Creator-Ausbau: Token-Rollen-UI (#3072, im Fokus). Übungs-Editor, CSV-Datei-Upload und Bild-Upload sind ausgeliefert (#3071).
  • EXP-024 — Schichtentrennung: Phase 2 (Plugin-Service-Module), Phase 3 (Frontend-U1).

Deferred (Phase 2 / community-finanziert)

  • EXP-029 — MED-10 (Partner-Onboarding-Doku): Business-Development- Artefakt, kein Code; Engpass ist die Partnergewinnung.
  • EXP-030 — Multi-User: MU-01..MU-05, MU-10, MU-20.
  • EXP-023 — Server-Features (Phase C): Community-Ratings, Trust 2, zentraler Index, Coach-Aggregation, Einmal-Invite-Tokens.
  • EXP-032 — CQV-04/05: Batch-LLM-Review mit Confidence-Score (teuer, Release-Lauf, nach CQV-01..03) + anonymer Fehler-Melden-Button → Content-Repo-Issue (moderiert, nach EXP-030).
  • EXP-027 — RTL-Infrastruktur: I18N-01 (RTL-Basis), I18N-04 (Arabisch-UI), I18N-05 (Koreanisch/Hangul).
  • EXP-013 Stufe 3 — KI-augmentierte adaptive Generierung (P-150..P-152).
  • Weitere Einzel-Posten: in ROADMAP.md > Detail (ANALYSIS-TARGET-DETECT-01, ANALYSIS-DOMAIN-SUGGEST-01, PLACEMENT-LANG-WARN-01, PT/TR/JA-Native-Review, SW-Cache-Knobs, iOS-Safari-Meta-Tags, Push-Notifications, native Wrapper) sowie in den P3-P5-Abschnitten unten (BADGE-EVAL-NPLUS1-01, PERF-EAGER-GLOBS-01, DEP-MYPY-2-01, DEP-ANTHROPIC-105-01, PLUGINFORGE-LIFECYCLE-UI-01, BACKEND-RADON-REMAINDER-01, BL-14..17, BL-19).

P3: when capacity allows

  • BACKEND-RADON-REMAINDER-01: opportunistic backend complexity burn-down for any module still above the radon warn band (cc > 15) now that the Phase 2 hard gate (#494/#495: blocks cc > 20, warns > 15) is live and .complexity-baseline is empty. No grandfathered offenders remain; this is for new warn-tier functions as they surface. Opportunistic.

  • BADGE-EVAL-NPLUS1-01: Badge evaluation fires one query per badge. evaluateBadgesForUser (frontend/src/storage/badges.ts) and badge_service.evaluate_user (gamification plugin) both load the catalog + earned rows (2 queries), then iterate ~28 evaluators where ~14 predicate / tier-metric helpers EACH run their own query, many re-scanning the SAME tables (sessions, lessonProgress, elementErrors, userXp) -> ~16-30 queries per evaluation. NOT on a page-load path (runs after lesson/session completion). Fix: build a shared metrics snapshot once and thread it through all predicates, in BOTH modes (keep the cross-language parity golden green). Cross-cutting refactor of 14+ functions x2 modes -> deferred. Filed from the 2026-06-03 performance audit (B-1 / C-1).

  • PLUGINFORGE-LIFECYCLE-UI-01: Consume v0.9.0 lifecycle visibility in Settings → Plugins. Backend half SHIPPED 2026-05-23: GET /api/plugins/inspect/{name} + api.plugins.inspect() API client + PluginInspection TypeScript type + 2 backend tests + 2 Vitest tests. The endpoint surfaces activated_at, last_config_change, source, filter_reason, load_error, version, target_application per plugin (404 on unknown name). Frontend half STILL PENDING: the scope estimate at filing time ("~40 LOC frontend, fold into the existing Settings → Plugins panel") assumed a Settings → Plugins panel exists. Re-audited Settings.tsx — no Plugins section currently exists in the UI. Building it from scratch is closer to 150-200 LOC (list active plugins via api.plugins.health() or manifests(), render a row per plugin with name + version + activated_at + source). Deferred until the next time Settings sees structural work, or someone reports needing the lifecycle info. The backend contract is stable and shipped; the panel can be built on top of it without further backend changes. Audit (docs/audits/pluginforge-0.9.0-adoption-signal-2026-05-21.md).

  • TOKEN-ALIAS-GATE-GAP-01: no-hardcoded-colors.test.ts blocks raw hex/rgb/hsl literals in consumer code, but does not see legacy alias tokens (var(--surface), var(--border), var(--danger), defined as pass-throughs to the canonical tokens in frontend/src/styles/legacy/00-head.css:87-97 — --surface to --bg-surface, --border to --border-primary, --danger to --error). An alias is already var()-bound, so the guard treats it as compliant; it verifies "no raw color", not "canonical token name", which is a different property than design-tokens.md's own "Legacy aliases" section implies it enforces. That section calls the aliases legacy and says to prefer the semantic names for new code, but that preference is a documented convention with no gate behind it anywhere in the guard suite (same shape as the earlier #2079-era gaps: a rule stated in prose, unchecked in CI). The aliases are alive and heavily used, not dead code: in frontend, grep -rnE "var\(--(border|surface|danger)\)" --include="*.tsx" --include="*.ts" src/components src/pages src/shared | grep -v "\.test\.ts" (consumer UI code, definitions and tests excluded) counted 138 call sites as of this filing, across exercise renderers, sync, settings and about-panel components — re-run before trusting the number in a later session. Migrating all of them is a large, visually-critical-surface change with no gate forcing it and no confirmed payoff — the cost of touching 138 sites plausibly exceeds the harm of the aliases surviving, so this item is the gate gap only, not a migration proposal. Whether migration ever pays for itself is a separate open question. Surfaced while auditing a UI-duplication extraction candidate: a repeated border-[var(--border)] bg-[var(--surface)] card wrapper at 4 sites (pages/content/RedeemInvite.tsx:110, pages/content/SetDeepLink.tsx:129, pages/content/AddRepo.tsx:164 and :183) — those 4 stay as the concrete example, not the scope of this item.

P4: deliberately deferred (future, SaaS)

  • BL-14: PostgreSQL migration — Replace SQLite with PostgreSQL for multi-user/SaaS deployment. Alembic migrations, connection pooling, Docker Compose with pg service.

  • BL-15: JWT authentication — User login, registration, password reset. Required for multi-user. Consider OAuth (Google, GitHub) as alternative.

  • BL-16: Multi-user support — Tenant isolation, shared curricula, team learning projects. Requires BL-14 + BL-15.

  • BL-17: Stripe integration — Premium plugins, subscription tiers. Requires BL-14 + BL-15 + BL-16.

  • ARCADE-SCORES-01: Personal best list per arcade game (idea 2026-09-24, pick up after EXP-052 slice 4). An arcade-score-store.ts next to ticket-store.ts and arcade-unlock-store.ts, same Dexie namespace pattern, showing the best five runs with date per game like an arcade cabinet's high-score screen. The four games have four notions of a score, so the record stores value AND direction per game: Snake counts up (score), Simon's value is the longest repeated sequence, Memory's completed two-card reveals count DOWN (fewer is better), TicTacToe has only an outcome (won/lost/draw), so a win/draw/loss tally or no entry. A leaderboard across people is out of reach: offline-first, data local in Dexie, no accounts, no shared state (same blocker as BL-15/BL-16). A shared ranking would also reach back into learning, because arcade play time comes from tickets earned by perfect lessons and streaks; a personal list does not. Hidden behind the game-mode and arcade switches, see #3216.

P5: speculative (no concrete trigger)

  • BL-19: Social features (share progress, study groups) — Requires multi-user. Far future.
  • EXP-035: Vercel AI Elements integration (AI-UI library, Tier 3 over custom). Codebase-audited; recommended first step is AIE-04 (Reasoning UI, net-new) using AI Elements as a pure UI layer (no Vercel AI SDK). AIE-03 (CodeBlock) + AIE-10 (Streaming) deprioritized — both already covered. See explorations/EXP-035-vercel-ai-elements.md.

Trigger-Gated Items

These activate when a specific condition is met:

Item Trigger
BL-07 Global subjects User creates 5th learning project
BL-11 PT/TR/JA First non-DE/EN/ES/FR/EL user request
BL-14..17 SaaS Decision to go commercial
BL-18..19 Social 100+ active users

Blocked / Upstream Wait

  • AUTH-03+ — cross-repo author publishing flow (EXP-025): the remaining author-companion scope beyond AUTH-01/02 (#529/#531, shipped) — publishing/maintaining a book companion repo across repositories. Needs shared-backend / cross-repo infrastructure (same dependency as the EXP-023 deferred set: community ratings, Trust 2, central index, coach aggregation). Deferred until that infrastructure lands.