argus
CLAUDE.md1 starsChanged 9 months agoMITView on GitHub
AI-powered multi-agent system: Slack Bot + Dashboard + Agent Orchestrator pnpm monorepo (@argus/ スコープ): 依存: slack-bot → agent-core, db, gmail, google-calendar, knowledge-personal, r2-storage, tiktok / dashboard → agent-core, db / orchestrator → agent-core, db, knowledge, gmail, google-calendar / knowledge → db / knowledge-personal → db / google-calendar → gmail / tiktok → db, gmail / gmail → db セッション設計・実行ループ・観測・Memory中心設計・権限分離の詳細は .claude/rules/architecture.md を参照 スキーマ定義: packages/db/src/schema.ts 主要テーブル: sessions, messages, tasks, knowledges, agents, agent_executions, lessons, sns_posts, inbox_tasks, daily_plans 詳細は .claude/rules/coding-conventions.md を参照 Dashboard files API…
- Reads credentials
In this repository24 other files it gives its agents.
- argus-workflowSkill
- daily-digestSkill
- daily-digest-thumbnailSkill
- doc-generateSkill
- image-generatorSkill
- knowledge-reportSkill
- mermaid-to-webpSkill
- podcast-builderSkill
- presentation-builderSkill
- session-summarySkill
- sns-github-publisherSkill
- sns-instagram-imageSkill
- sns-note-writerSkill
- sns-podcast-creatorSkill
- sns-publisherSkill
- sns-qiita-writerSkill
- sns-threads-posterSkill
- sns-tiktok-creatorSkill
- sns-x-posterSkill
- sns-youtube-creatorSkill
- sns-zenn-writerSkill
- tts-dictSkill
- ttsSkill
- video-studioSkill
The file
Files
Argus
AI-powered multi-agent system: Slack Bot + Dashboard + Agent Orchestrator
Quick Reference
pnpm dev # 全アプリ並列起動(build後)
pnpm test # 全テスト実行
pnpm build # 全パッケージビルドArchitecture
pnpm monorepo (@argus/ スコープ):
| レイヤー | コンポーネント | 役割 |
| Frontend | apps/dashboard (Next.js 16, Port 3150) | セッション監視、ナレッジ管理UI |
| Backend | apps/agent-orchestrator (Port 3950) | エージェント実行・スケジューリング |
| Backend | apps/slack-bot (Port 3939) | Slack連携・メッセージ処理 |
| Core | packages/agent-core | Claude SDK ラッパー(session + hooks + text-utils) |
| Core | packages/db | Drizzle ORM (PostgreSQL) |
| Knowledge | packages/knowledge | Knowledge MCP Server |
| Knowledge | packages/knowledge-personal | Personal Knowledge MCP Server |
| Integration | packages/gmail | Google API(OAuth2 + Gmail) |
| Integration | packages/google-calendar | Google Calendar MCP Server |
| Integration | packages/r2-storage | Cloudflare R2 ストレージクライアント |
| Integration | packages/tiktok | TikTok API + PKCE 認証 |
| Config | .claude/ | ルール / skills / permissions |
依存: slack-bot → agent-core, db, gmail, google-calendar, knowledge-personal, r2-storage, tiktok / dashboard → agent-core, db / orchestrator → agent-core, db, knowledge, gmail, google-calendar / knowledge → db / knowledge-personal → db / google-calendar → gmail / tiktok → db, gmail / gmail → db
セッション設計・実行ループ・観測・Memory中心設計・権限分離の詳細は .claude/rules/architecture.md を参照
.claude/ ディレクトリ
skills/— スキル定義(SKILL.md)。ls .claude/skills/で一覧確認rules/— architecture, coding-conventions, context-engineeringagents/— analyzer, code-reviewer, collector, executorsettings.json— 権限設定(deny-first)- 作業:
.claude/agent-workspace/(一時) / 納品:.claude/agent-output/YYYYMMDD-*/(永続)
データモデル
スキーマ定義: packages/db/src/schema.ts
主要テーブル: sessions, messages, tasks, knowledges, agents, agent_executions, lessons, sns_posts, inbox_tasks, daily_plans
Tech Stack
- Node.js >= 22.12.0, pnpm 10.x, TypeScript 5.x (strict, ESM)
- Next.js 16, React 19, Tailwind CSS 4
- Drizzle ORM + postgres.js, Supabase PostgreSQL
- Vitest 4, Testing Library
- Deploy: ローカルMac + PM2 + Cloudflare Tunnel/Access
Conventions
詳細は .claude/rules/coding-conventions.md を参照
- ESM統一 —
import/export、node:プレフィックス、パッケージ内は.js拡張子付きimport - ファイル名: kebab-case / コンポーネント: PascalCase / DBカラム: snake_case
- エラーハンドリング:
success: booleanフラグで返す(throwしない) - テスト: ソースと同ディレクトリにコロケーション (
foo.ts+foo.test.ts) - 環境変数: ルート
.envに一本化 - マイグレーション:
drizzle-kit generate --customで意味のある名前を付ける(例:add-session-indexes)
メディア配信
Dashboard files API (/api/files/[...path]) で MP4/MP3 を配信。
DASHBOARD_BASE_URL (デフォルト: http://localhost:3150) + /api/files/{output-dir}/output.mp4
Slack にはファイルパスではなく URL を貼る。
Deploy
ローカルMac(PM2プロセス管理)+ Cloudflare Tunnel + Access(メール認証) 詳細: docs/DEPLOYMENT.md
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

