Previous code continuously forced activeTab to 'messages' when not
authenticated, making dashboard and voice tabs unclickable.
Now uses useEffect once on mount to reset localStorage carryover,
then lets tab switching work normally.
tailwind.config.js overrides spacing.xs to 4px, which also overrides
max-w-xs to max-width: 4px instead of the default 20rem (320px).
Changed to explicit max-w-[280px] to bypass the corrupted token.
When a user visited the Live tab in a previous session, localStorage
stored activeTab: 'live'. On page reload, even with the new auth guard,
the app would try to render the Live tab and show the auth overlay.
Now if not authenticated, activeTab is always forced to 'messages'.
Fixes MascotChatbot messages wrapping 1 character per line.
overflow-wrap:anywhere handles long unbroken strings better
than break-words which only breaks at word boundaries.
Messages, analysis, mascot-chat, recordings, ui-state, and guilds
endpoints are now accessible without authentication.
Only voice (connect/disconnect/transmit) and media (queue/skip/stop/volume)
remain behind the admin password guard.
Cloudflare Rocket Loader transforms <script type="module"> into
type="<hash>-module", which the browser cannot parse as ES module.
The React app never mounts — blank page on every route.
Fix: Vite plugin adds data-cfasync="false" attribute to the module
script tag, telling Rocket Loader to skip transformation.
Also: remove dist/ from the commit (build artifacts excluded).
- Replace @apply im-btn in im-btn-primary/secondary/ghost
with inline Tailwind utilities + CSS properties
- Replace @apply im-status-dot in status dot variants
with inline properties
- Tailwind 4 cannot @apply component-level classes from within
@layer components — only standard utility classes are valid
Two-layer fix for forwarded messages showing as empty/clean:
Layer 1 (messageMetadata.ts): getReferencedMessageContent() now falls
back to message.messageSnapshots Collection when channel.messages.cache
lookup fails. Discord stores forward content in message_snapshots API
field, not in message.content.
Layer 2 (moderationBuilders.ts): buildReferenceXml() now parses msg.
metadata JSON to extract reference.content when DB getMessageById()
fails (cross-server forwards not in local DB).
Previously: forward messages captured with empty parentContent →
LLM saw no reference text → '99% confidence, pesan kosong'.
Now: forward content flows through capture → metadata → analysis.
- SELECT transcription FROM voice_recordings fails because table has no such column
- Removed non-existent transcription from SELECT and interface
- Previous sql.raw() parameter binding also broken — fixed with proper drizzle sql``
- Hand-rolled , params with sql.raw() didn't actually bind values
- Drizzle's sql.raw() just inserts literal text — no parameter binding
- Replaced with proper sql`` tagged templates + sql.join() for conditions
- Each filter now correctly binds via drizzle's parameterized query
- Previously: text+media message went ONLY to media array → text waited for vision
- Now: text part goes to text batch (immediate LLM analysis), media parallel
- DB update is idempotent — second write to same message_id overwrites
- User sees text moderation results instantly, media follows when ready
- text-only and media analysis now run concurrently via Promise.all
- text no longer blocks on media download + vision analysis
- each path independently saves to DB when its own results are ready
- same batch still uses single context fetch + attachment lookup
- Remove shouldSearchContent() trigger gate — search runs on all messages
- extractSearchQueries() now extracts from ANY message, not just trigger-matched
- Redis cache (24h TTL) prevents redundant searches for same query
- initSearxngCache() lazy-connects via config.REDIS_URL
- Cache miss→API, hit→skip — fire-and-forget writes
- Both text batch + media path simplified
- Remove hardcoded hentai title lists (Boku no Pico, Euphoria, etc.) from prompt
- Remove hardcoded SARA examples from prompt (Kitabonia, etc.)
- Prompt now tells LLM to use <web_searches> as evidence instead of hardcoded knowledge
- Evidence priority: searxng > web_content > media_analysis > internal model
- Code-side triggers in searxngSearch.ts still filter which messages to search
- LLM makes final decision based on search results, not static lists
- Video frame extraction via ffmpeg (4 key frames per video → vision LLM)
- Video display in FE MessageCard with HTML5 <video> player
- Reply/forward/crosspost indicator in FE + pipeline in DG/BE
- Fix: missing sanitizeAiContent + escapeXml in media path (prompt injection)
- Optimize: text-only batch results saved to DB immediately, no longer wait for media analysis
- BE mapper/schema/repo: add reference fields (is_reply, is_forward, etc.)
- userProfileLearner.ts: filter query to only clean messages (eq ai_status='clean')
to prevent profile contamination from flagged content. Also select channel_id
to group messages by channel in prompt, enabling channel-aware personality
summarization (user may behave differently across channels).
- llmModerationClient.ts (runSimpleTextFallback): inject user profile into
both the classify prompt and the reason prompt, so even the last-resort
fallback path has personality/memory context instead of being blind.
Migration SQL file existed but was never registered in _journal.json,
so Drizzle skipped it. Every voice recording insert failed with:
PG code 42703: column "transcription" does not exist
Co-Authored-By: Claude <noreply@anthropic.com>
Drizzle wraps PG errors — the previous log only showed the formatted query
text. Now extracts code, detail, schema, table, constraint, severity from
the error envelope and from .cause to make the actual Postgres error visible.
Co-Authored-By: Claude <noreply@anthropic.com>
asepharyana