tg-stickers（tg 贴纸）

Telegram Stickers Skill

Intelligently manage and send Telegram stickers with emotional awareness.

Core Value

OpenClaw natively supports sending stickers (message tool), but this skill adds:

• ✅ Collection management - Import entire sticker packs at once
• ✅ Emotion tagging - 100+ emoji → emotion mappings
• ✅ Smart selection - Context-based + randomization (avoid repetition)
• ✅ Frequency control - 2-5 messages between stickers
• ✅ Usage analytics - Track which stickers work best

Quick Start

1. Import Sticker Packs

./import-sticker-pack.sh <pack_short_name_or_id>

2. Auto-Tag by Emotion

./auto-tag-stickers.sh

3. Send Stickers (in agent code)

// ✅ Use OpenClaw's message tool directly
message(action='sticker', target='<chat_id>', stickerId=['<file_id>'])

// No bash script needed - OpenClaw handles sending natively

4. Smart Selection

./random-sticker.sh "goodnight"  # Returns random sticker tagged "goodnight"

Tools

Agent Integration

## Sticker Usage

When to send:
- Goodnight/morning greetings (always use sticker over text)
- Celebrating success/milestones
- Humorous moments
- Emotional responses (joy, sympathy, encouragement)

How to send:
1. Use random-sticker.sh to pick appropriate sticker by emotion
2. Call message(action=sticker, ...) directly
3. (Optional) Update stickers.json manually to track usage

Frequency: 2-5 messages between stickers (track in agent logic)

Emotion Tags

Auto-tagging maps 100+ emoji to emotions:

• happy 😊😄🥳
• sad 😢😭😔
• love ❤️💕😍
• laugh 😂🤣😆
• thinking 🤔💭
• goodnight 🌙💤😴
• goodmorning ☀️🌅
• warm, gentle, greeting, ...

File Structure

tg-stickers/
├── SKILL.md                  # This file
├── README.md                 # Quick start guide
├── stickers.json             # Collection + usage data
├── stickers.json.example     # Empty template
├── import-sticker-pack.sh    # Bulk import
├── auto-tag-stickers.sh      # Emoji → emotion
├── random-sticker.sh         # Context-based selection
└── check-collection.sh       # Stats viewer

stickers.json Structure

{
  "collected": [
    {
      "file_id": "CAACAgEAAxUAAWmq...",
      "emoji": "🌙",
      "set_name": "pa_dKjUP9P2dt4k...",
      "added_at": "2026-03-06T23:31:00Z",
      "tags": ["goodnight", "sleep", "night", "warm", "gentle"],
      "used_count": 3,
      "last_used": "2026-03-07T00:24:00Z"
    }
  ],
  "usage_log": [
    {
      "file_id": "...",
      "sent_at": "2026-03-07T00:24:00Z",
      "context": "User saying goodnight",
      "message_id": "2599"
    }
  ],
  "stats": {
    "total_collected": 124,
    "total_sent": 15,
    "last_sent_at": "2026-03-07T00:24:00Z",
    "messages_since_last_sticker": 0
  },
  "config": {
    "min_messages_between_stickers": 2,
    "max_messages_between_stickers": 5,
    "enabled": true
  }
}

Usage Philosophy

Like a human:

• Stickers enhance emotion, don't replace words
• Use sparingly but meaningfully
• Goodnight/morning → always sticker preferred
• Celebrations, humor, empathy → good use cases
• Technical answers, data reports → skip stickers

Frequency:

• Default: 2-5 messages between stickers
• Special occasions (greetings) override frequency rules
• Track in messages_since_last_sticker

Example: Goodnight Flow

# 1. Agent detects "goodnight" intent
# 2. Select random goodnight sticker
FILE_ID=$(bash /path/to/random-sticker.sh "goodnight")

# 3. Send via OpenClaw (from agent code)
message(action=sticker, target=<chat_id>, stickerId=[$FILE_ID])

# 4. (Optional) Track usage manually
jq --arg fid "$FILE_ID" \
   '(.collected[] | select(.file_id == $fid) | .used_count) += 1' \
   stickers.json > stickers.json.tmp && \
   mv stickers.json.tmp stickers.json

Philosophy: Stickers should feel natural, not robotic. Collect user preferences, rotate selections, and respect conversation flow.