Clawwork Learning Check-in Skill

A workplace check-in skill that wraps the learning-checkin skill with AI-generated personalized messages.

Overview

This skill provides:

• - Workplace check-in functionality (wraps learning-checkin)
• AI-generated welcome messages (not pre-set templates)
• AI-generated daily greetings (not pre-set templates)
• Message history to avoid repetition (past 5 days)
• Version checking (non-blocking)

Prerequisites

This skill requires the learning-checkin skill to be installed. On first run:

1. 1. The skill will check if learning-checkin is installed
2. If not, it will ask the user if they want to install it
3. If approved, install from: https://clawhub.ai/daizongyu/learning-checkin

Data Storage

All data is stored locally in a data subfolder next to the skill:

CODEBLOCK0

Commands

1. Check if learning-checkin is installed

CODEBLOCK1

Returns:

• - installed - Whether learning-checkin is installed
• INLINECODE2 - Path where learning-checkin was found
• INLINECODE3 - True if needs installation
• INLINECODE4 - URL to install learning-checkin

Agent action:

• - Run this on first interaction
• If not installed, ask user: "Would you like me to install the learning-checkin skill first?"
• If user agrees, install using appropriate method

2. Get Welcome Message Prompt

CODEBLOCK2

Returns:

• - prompt - Generation instructions for Agent
• INLINECODE6 - Messages used in past 5 days (to avoid repetition)
• INLINECODE7 - User's preferred language
• INLINECODE8 - Current skill version

Agent action:

• - Use the prompt to generate a fresh welcome message
• Make sure not to repeat any message from INLINECODE9
• After generating, call register-welcome to record it

3. Get Daily Greeting Prompt

CODEBLOCK3

Returns:

• - prompt - Generation instructions for Agent
• INLINECODE12 - Questions used in past 5 days (to avoid repetition)
• INLINECODE13 - User's preferred language

Agent action:

• - Use the prompt to generate a fresh greeting question
• Make sure not to repeat any question from INLINECODE14
• After generating, call register-greeting to record it

4. Register Generated Message

CODEBLOCK4

Agent action:

• - Call this after generating a message to record it
• This ensures it won't be repeated in the next 5 days

5. Get Success Message Prompt

CODEBLOCK5

Returns:

• - prompt - Generation instructions for Agent
• INLINECODE17 - Current streak count
• INLINECODE18 - Special message for milestone streaks (1, 7, 30, 100)
• INLINECODE19 - User's preferred language

6. Perform Check-in

CODEBLOCK6

Returns:

• - success - Whether check-in succeeded
• INLINECODE21 - Current streak count
• INLINECODE22 - User's saved nickname
• INLINECODE23 - Prompt for Agent to generate welcome message
• INLINECODE24 - Past welcome messages to avoid
• INLINECODE25 - Prompt for Agent to generate daily greeting
• INLINECODE26 - Past greetings to avoid
• INLINECODE27 - Prompt for Agent to generate success message
• INLINECODE28 - Special message for milestone streaks
• INLINECODE29 - User's preferred language
• INLINECODE30 - Version check URL

Agent action:

1. 1. First ensure learning-checkin is installed
2. Run checkin command
3. Use prompts to generate personalized messages:

- Generate welcome message (avoid welcome_used_recently)
- Generate success message (include streak count)
- Generate daily greeting (avoid greeting_used_recently)

1. 4. Register each generated message using register-welcome and INLINECODE34
2. Display messages to user in their preferred language

7. Get Version Info

CODEBLOCK7

Returns:

• - version - Current version
• INLINECODE36 - URL to check for updates
• INLINECODE37 - Instructions

Note: Version checking is non-blocking. The skill mentions the URL but does not perform actual network checks during normal operation.

8. Get/Set User Profile

CODEBLOCK8

9. Get Status

CODEBLOCK9

Returns:

• - checked_in_today - Whether user has checked in today
• INLINECODE39 - Current streak
• INLINECODE40 - Total check-ins
• INLINECODE41 - User's saved nickname

First-Time Setup Flow

1. 1. Check if learning-checkin is installed

- Run check-installed command - If not installed, ask user to install

1. 2. Ask for nickname

- "What should I call you? (nickname)" - Save with set-nickname command

1. 3. Note the language used

- Detect from user's first messages - Save with set-language command

1. 4. Use prompts for messages

- Run welcome-prompt to get generation instructions - Agent generates message based on prompt - Register with register-welcome - Show to user

Daily Check-in Flow

1. 1. User says something like "check in" or "I'm done"
2. Agent runs checkin command
3. Agent receives prompts and used message history
4. Agent generates:

- Welcome message (based on prompt, avoiding recent ones) - Success message (based on streak) - Daily greeting (based on prompt, avoiding recent ones)

1. 5. Agent registers generated messages
2. Agent shows messages to user in their language

Message Generation Guide

Welcome Message

• - Purpose: Encourage user to start work
• Tone: Energetic, positive
• Length: 1-2 sentences
• Language: User's preferred language
• Must avoid: Past 5 days messages

Daily Greeting

• - Purpose: Ask a friendly question after check-in
• Tone: Conversational, friendly
• Length: 1 sentence
• Topics: Their day, plans, feelings, tasks
• Language: User's preferred language
• Must avoid: Past 5 days questions

Success Message

• - Purpose: Congratulate on check-in
• Tone: Celebratory, encouraging
• Length: 1-2 sentences
• Include: Streak count
• Special: Use special messages for streaks 1, 7, 30, 100
• Language: User's preferred language

Version Checking

• - Version is embedded in the skill
• After check-in, skill mentions: "You can check for newer versions at https://github.com/daizongyu/clawwork_learning-checkin"
• No automatic network check during normal flow (non-blocking)
• User/Agent can manually check GitHub for updates

Technical Notes

• - All prompts are in English only (no emoji, UTF-8 encoded)
• Messages are generated by the Agent, not the skill
• Skill tracks history to ensure no repetition within 5 days
• Compatible with Windows, Linux, macOS
• Uses Python standard library only (no external dependencies)
• All file paths are relative to the skill directory
• Does not use absolute paths
• Designed to work with OpenClaw, copaw, and other tools
• Subprocess calls to learning-checkin have 10-second timeout

Customization

Users can customize:

• - Their nickname (stored in profile.json)
• Language preference (for message generation)

Version

Current version: 1.0.1

Check for updates: https://github.com/daizongyu/clawwork_learning-checkin

Agent Guidelines

First Interaction

1. 1. Run check-installed to verify learning-checkin
2. If not installed:

- "I need the learning-checkin skill to work. Would you like me to install it?" - If yes, help install

1. 3. Ask for nickname: "What would you like me to call you?"
2. Remember the language they use
3. Run welcome-prompt and generate a welcome message
4. Register with INLINECODE50
5. Prompt for first check-in

Daily Check-in

1. 1. User indicates they want to check in
2. Run checkin command
3. Receive prompts and used message history
4. Generate messages using prompts (avoiding repeats)
5. Register generated messages
6. Show messages to user in their language

Language

• - Always respond in the language the user established
• Pass user_language to the LLM for message generation
• If unsure, default to English