- Add explicit system rule that IMPHNEN is the project's own name, not religion
- Rename 'Imphnemia 11:17' example to 'Kitabonia 11:17' to avoid name collision
- Ensures mentioning/promoting the project URL is not flagged as SARA
Co-Authored-By: Claude <noreply@anthropic.com>
Add user_profiles table, store, and background learner worker
that summarizes user communication style, topics, and personality.
- New user_profiles table (user_id PK, guild_id, profile_summary, last_analyzed_at)
- userProfileStore.ts — CRUD (get/update) following channelCultureStore pattern
- userProfileLearner.ts — background worker: queries 100 recent msgs per user,
calls LLM for personality summary, updates every 12h
- Inject <user_profile> XML tag per-message in moderation prompt
- Start worker alongside cultureLearner in aiAnalyzer.ts
- Migration 0008 for user_profiles table
Co-Authored-By: Claude <noreply@anthropic.com>
Update the moderation prompt to include high-priority detection categories for:
- Fake scripture/verse parodies
- Claims of divinity or false religious movements
- Misuse of theological terms as internet slang/memes
- Mockery of religious figures and rituals
This change ensures stricter enforcement of SARA (Suku, Agama, Ras, Antargolongan) policies by explicitly defining religious blasphemy and parody as high-severity violations.
user_history (riwayat flag sebelumnya) dan clean_streak/total_infractions
dikirim ke LLM setiap kali menganalisis pesan — ini bikin self-fulfilling
prophecy: user yg pernah kena false positive jadi makin gampang dituduh
lagi, dan link Instagram pun dianggap sexual_deviation cuma karena
riwayat user.
Changes:
- Hapus getUserRecentInfractions dari text batch path
- Hapus getUserRecentInfractions dari media analysis path
- Hapus import getUserRecentInfractions yg gak dipakai
- Ubah instruksi prompt dari 'jadilah lebih tegas jika riwayat jelek'
jadi 'setiap pesan dinilai berdasarkan isinya sendiri'
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Common Indonesian religious expressions like 'Astaghfirullah', 'Astaga',
'Alhamdulillah', 'Subhanallah', dll were being flagged as vulgar_language
by the LLM. Added explicit rule that these are normal religious/cultural
expressions in Indonesia - not vulgar language - even in all-caps or
with repeated letters.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Adds explicit rule that Indonesian names/nicknames like 'Sapik' (Syafik),
'Ayang', 'Dek', 'Bang', 'Mas', etc. are NOT furry or sexual_deviation references
- Adds rule prohibiting the LLM from inventing slang meanings for words
it doesn't recognize - default to innocent until proven guilty
- Prevents false positive cascade where LLM confuses names with furry slang
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Nama karakter game/anime populer seperti 'Furina' dari Genshin Impact
sering kena false positive sebagai 'sexual_deviation' karena kemiripan
fonetik dengan kata 'furry'. Menambahkan aturan eksplisit bahwa nama
karakter fiksi normal bukan referensi furry fetish, dgn pengecualian
jika konteks pesan secara eksplisit membahas aspek fetish/seksual.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Update moderation guidelines to prevent false positives on song lyrics, poems, memes, and literary quotes, ensuring political or revolutionary content is not flagged as conflict instigation unless accompanied by explicit incitement.
- Add automatic fallback to streaming mode if the provider rejects non-streaming requests with a 400 error
- Refactor `llmChat` to use an internal execution function to support retry logic with modified parameters
- Update moderation prompt to explicitly allow Japanese pop culture terms (e.g., "moe", "waifu", "wibu") to prevent false positive sexual deviation flags
Implements a context-aware moderation system by tracking user behavior
and channel-specific norms to improve AI decision-making accuracy.
- Adds `user_reputations` table to track trust scores, clean streaks,
and infraction history.
- Adds `channel_cultures` table to store AI-generated summaries of
channel-specific norms and slang.
- Implements `userReputationStore` to autonomously update user scores
based on moderation outcomes (clean vs. flagged).
- Implements `cultureLearner` and `channelCultureStore` to manage
evolving channel contexts.
- Enhances LLM prompts to inject user reputation (trust scores,
history) and channel culture summaries, enabling "wisdom-based"
moderation (e.g., giving benefit of the doubt to high-trust users).
- Integrates reputation and culture updates into the existing
`aiAnalyzer` pipeline.
Refactors the AI moderation pipeline to improve concurrency control and
cache efficiency by moving from user-centric to content-centric caching.
- Implements a distributed locking mechanism for media analysis using
`acquireMediaAnalysisLock` to prevent redundant LLM vision calls across
multiple pods.
- Transitions text moderation caching from `user_mod:userId:hash` to a
purely content-based `text_mod:hash` approach to increase hit rates.
- Enhances `getPendingMessagesByConversation` with atomic transactions
and `FOR UPDATE SKIP LOCKED` to safely transition messages from
`pending` to `processing` state.
- Adds `processing` status to the `AIStatus` type and database schema to
track active analysis lifecycles.
- Implements polling logic in `llmModerationClient.ts` to wait for
in-progress media analyses.
- Instructs the AI to decode combinations of regional indicator emojis (e.g., 🇬 🇦 🇾) and custom letters spelling out words, rather than dismissing them as 'just a series of emojis'
- Added a specific few-shot example (Contoh 15) to demonstrate flagging this technique when used to spell banned words
- Added explicit zero-tolerance rule for anatomical/sexual vulgarity (e.g. titten, kontol), explicitly forbidding the AI from passing them off as 'casual conversation' or 'jokes'
- Expanded sexual_deviation rule to explicitly cover brief mentions of BL (Boys Love), yaoi, yuri, and LGBT topics, instructing the AI to flag them regardless of casual context
Sebelumnya: 'Pesan hanya berisi attachment tanpa teks yang melanggar'
Sekarang: harus deskriptif berdasarkan tipe konten:
- Text only: '[user] membahas tentang <topik>. <konteks>.'
- Image only: 'Gambar berupa <jenis>. Terlihat <isi>.'
- Text+Image: '[user] mengirim <gambar> sambil membahas <topik>.'
Tambahkan contoh baik vs buruk di OUTPUT_INSTRUCTIONS sebagai format wajib.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Sebelumnya aturan 'percaya teks terlebih dahulu' membuat model abaikan
deskripsi gambar saat teks kosong. Semua image-only message di-clean.
Fix:
- SYSTEM_RULES: pisah Mode 1 (teks+gambar) dan Mode 2 (hanya gambar)
- Mode 2: deskripsi gambar jadi bukti utama, WAJIB dibaca
- Gambar terminal/chat/editor kode/casual → clean
- Gambar dengan elemen judi NYATA (chip, roulette, odds) → flag
- 2 contoh few-shot baru: terminal clean, situs judi flag
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Root cause: vision model diminta untuk 'flag' dan 'menilai' gambar,
sehingga screenshot terminal/chat biasa diklaim sebagai 'situs perjudian'.
Fix:
- vision prompt: HANYA deskripsi objektif (objek, teks, layout, jenis gambar)
- larang tegas kata 'gambling', 'judi', 'pelanggaran', 'harus dihapus'
- tambah buildGeneralImageVisionPrompt di discord-gateway stickerPrompt
- MEDIA_INSTRUCTIONS: tegaskan batch LLM adalah hakim, vision hanya saksi mata
- deskripsi netral (terminal, chat, editor kode) tidak boleh jadi dasar flag
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
- Extract services into services/{frontend,backend,discord-gateway}
- Create packages/shared/ for shared logger, errors, utils, types
- Setup Modular MVC pattern in backend (controller→service→repository)
- Setup event-driven architecture in discord-gateway with Redis pub/sub
- Move Docker files to infra/docker/ with per-service Dockerfiles
- Update docker-compose.yml to use Traefik-only routing (no port exposes)
- Update GitHub Actions deploy workflow for multi-service matrix build
- Fix all import paths and resolve type errors across all services
- All 3 services pass tsc --noEmit clean
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
