Commit Graph
11 Commits
Author SHA1 Message Date
MythEclipse 1c46a8a084 feat(ai-moderation): add conflict instigation, offensive username, and discrimination moderation flags and prompt rules 2026-06-04 12:49:24 +07:00
MythEclipse c1de1279a5 There are no staged changes to commit. The staging area is empty — git status shows a clean working tree with nothing staged. 2026-06-02 23:07:11 +07:00
MythEclipse 1f3d1ac3f4 API Error: API returned an empty or malformed response (HTTP 200) — check for a proxy or gateway intercepting the request 2026-06-02 23:06:55 +07:00
MythEclipse 774472b6ba perf(discord-gateway): add tiktoken token counting, vision LRU cache, PromptMode few-shot splits, and Piscina maxThreads config 2026-06-02 22:50:53 +07:00
MythEclipse f163ead3cf feat(discord-gateway): add local badword pre-filter, accurate token estimation, Piscina maxThreads config, and PromptMode support 2026-06-02 22:50:23 +07:00
MythEclipse 6d2da7ce1d perf(discord-gateway): start AI moderation analysis immediately, running attachment uploads in parallel instead of blocking 2026-06-02 22:49:57 +07:00
MythEclipseandClaude Opus 4.8 62d7d19b53 fix: analysis output wajib deskriptif — bukan generic placeholder
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>
2026-06-02 11:24:21 +07:00
MythEclipseandClaude Opus 4.8 a69de85564 fix: prompt dua mode — gambar+teks vs gambar saja
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>
2026-06-02 11:15:53 +07:00
MythEclipseandClaude Opus 4.8 55ee8ade16 fix: vision model hanya deskripsi, tidak memutuskan moderasi
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>
2026-06-02 11:08:41 +07:00
MythEclipseandClaude Opus 4.8 f1ddca5eee fix: false positive gambling detection + route collision + WS events (#1)
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-02 10:44:27 +07:00
MythEclipseandClaude Opus 4.8 c48a0c5e3b refactor: split monolith into 3 microservices (frontend, backend, discord-gateway)
- 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>
2026-06-01 21:44:29 +07:00