メニュー
Process Telegram messages via GramJS MTProto client. Read messages, generate AI drafts, save drafts, and create intro groups. Use when handling Telegram conversations or facilitating introductions.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/Gerstep/cybos --skill telegramコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
Guide founders and teams through a structured positioning statement exercise. Combines competitive research, strategic questioning, and team alignment to produce a sharp, honest positioning statement. Use when a company needs to define who they are, what they offer, and why anyone should care.
Iterative critique and improvement of long-form content (guidebooks, playbooks, essays). Launches parallel judge subagents for multi-dimensional critique, synthesizes findings, presents proposals for user approval. Never edits without consent.
Senior UI/UX Engineer. Architect digital interfaces overriding default LLM biases. Enforces metric-based rules, strict component architecture, CSS hardware acceleration, and balanced design engineering.
SOC 職業分類に基づく
SKILL.md を表示中
| name | telegram |
| description | Process Telegram messages via GramJS MTProto client. Read messages, generate AI drafts, save drafts, and create intro groups. Use when handling Telegram conversations or facilitating introductions. |
Process Telegram messages via GramJS MTProto client. Supports unread messages, specific users, message requests, and group creation for introductions.
CRITICAL: NEVER SEND MESSAGES. Only save drafts.
| Capability | Description |
|---|---|
| Unread Mode | Process N unread conversations |
| User Mode | Find specific person by username/name (any read state) |
| Requests Mode | Process message requests folder (non-contacts) |
| Entity Context | Load context from database for known contacts |
| Draft Replies | AI generates contextual reply drafts |
| Save Drafts | Save drafts to Telegram (no sending) |
| Create Groups | Create new groups and add members for introductions |
| Dialog Cache | Fast lookups via cached dialog IDs (~1s vs ~90s) |
| Mark Unread | Re-mark conversations as unread after processing |
| History | Save per-person history to vault context/telegram/ |
workflows/process-messages.md - Full workflow for reading/replyingworkflows/create-intro-group.md - Workflow for facilitating introductions| Script | Purpose |
|---|---|
scripts/telegram-gramjs.ts | GramJS MTProto client - fetch messages, populate cache |
scripts/telegram-save-drafts.ts | Save AI drafts to Telegram from work file |
scripts/save-telegram-draft.ts | Quick draft save by @username (instant) or name (dialog search) |
scripts/telegram-create-group.ts | Create groups and add members for intros |
# Unread mode (default)
/cyber-telegram # 1 unread dialog
/cyber-telegram --count 3 # 3 unread dialogs
# User mode
/cyber-telegram --user "@username" # By username
/cyber-telegram --user "Name" # By name
# Requests mode
/cyber-telegram --requests # Message requests folder
# Modifiers
/cyber-telegram --dry-run # Read only
/cyber-telegram --no-mark-unread # Don't preserve unread state
# Create intro group
bun scripts/telegram-create-group.ts --dry-run --title "A <> B" --users "@user1,@user2"
bun scripts/telegram-create-group.ts --title "A <> B" --users "@user1,@user2" --draft "Intro message"
See workflow files for full documentation.
Dialog cache (~/.cybos/telegram/dialog-cache.json) stores all dialog IDs with access hashes for fast lookups:
telegram-save-drafts.ts to avoid fetching all dialogssave-telegram-draft.ts uses fast contacts.ResolveUsername API for @handles — resolves any Telegram user/channel instantly without scanning dialogs. Falls back to searching recent dialogs (top 200) for name-based lookups.
Drafts only - never sends messages automatically. User reviews and sends manually in Telegram.