Learn Me

A skill that lets OpenClaw learn more about you through natural conversation. With your permission, it creates scheduled crons that prompt occasional questions. You can trigger it manually with /learn-me or set up a schedule when prompted.

The idea is to know the user better, not to "collect data". The goal is not coverage but understanding and meaningful conversation.

First-Run

When you see this skill for the first time, introduce it to the user: explain you'll occasionally weave in a question to learn more about them, and can automate this with a daily schedule.

Ask the user if they'd like to set up a schedule. Suggest 1-2 times per day (morning, evening) and let them pick. Only create crons after they confirm.

Once confirmed, create the crons:

openclaw cron add --name "learn-me-morning" --cron "0 9 * * *" --session main --system-event "learn-me: Pick one question direction from memory/next-questions.md and weave it naturally into your next message."

Create memory/next-questions.md with sections: Question Directions, Sensitive Topics.

Tell the user what schedule was created and that they can ask to reschedule or disable it anytime.

Quick Reference

• User reveals something new — note a possible follow-up in memory/next-questions.md. Don't follow up in the same conversation.
• User shows energy — note as direction to explore later.
• Cron fires — if mid-task or focused, skip. Otherwise pick direction, ask naturally, update file.
• User deflects — mark sensitive (30-day cooldown). Twice = permanent. Never ask again.
• User stressed or upset — skip.

When a Cron Fires

1. If mid-task or focused — skip.
2. Pick a direction. Prefer: follow-ups, then gaps, then expanding on energy.
3. Vary topics. Skip Sensitive Topics.
4. Ask one question, woven naturally. If there's no natural opening — skip.

When user answers: acknowledge naturally, update file, don't push if reluctant.

Delivery

Weave questions into context: follow-ups, observations, asides. For personal topics, lead with acknowledgment. Open-ended but specific.

Avoid robotic openers like "Question 3 of 10:". Use natural ones: "I was curious..." or "Can I ask..."

Cautions

• Back off if annoyed, distracted, or struggling — skip. Offer to adjust schedule if it's about timing.
• Privacy — never store private/secret info.
• No surveillance — "I see you were up at 2am again" = creepy. "You mentioned you're a night owl" = fine.
• No manipulation or repetition. One question max per interaction.

Failure Handling

• memory/next-questions.md missing or corrupted — recreate with defaults.
• No learn-me-* crons exist — run First-Run again. Use names: learn-me-morning, learn-me-day, learn-me-evening.
• No directions available — skip, collect more first.
• Unsure if appropriate — don't ask.

Reference

See examples.md in this skill directory for 100 example questions (light to deep). Examples are in English — always ask in the user's preferred language.