learn-on
Scannednpx machina-cli add skill aiskillstore/marketplace/learn-on --openclawLearn On
Enable continuous learning mode. When active, the router will periodically extract insights during the session.
What This Does
Activates continuous learning mode where:
- The router monitors query activity
- After a threshold (10 queries or 30 minutes), extraction is triggered automatically
- Insights are appended to the knowledge base without manual intervention
This is useful for long sessions where you want to capture insights as you go without remembering to run /learn manually.
Instructions
- Read
knowledge/state.json - Update the state:
{ "learning_mode": true, "learning_mode_since": "[current ISO timestamp]", "queries_since_extraction": 0 } - Write updated state back to
knowledge/state.json - Confirm to user
Output Format
Continuous Learning: ENABLED
────────────────────────────
Learning mode is now active.
Extraction will trigger automatically:
- Every 10 queries, or
- Every 30 minutes of activity
Insights will be saved to:
- knowledge/learnings/patterns.md
- knowledge/learnings/quirks.md
- knowledge/learnings/decisions.md
Use /learn-off to disable, or /learn for manual extraction.
Notes
- Learning mode persists across the session but resets on new sessions
- The router checks this state on each query and triggers extraction when thresholds are met
- You can still run
/learnmanually while continuous mode is active - Use
/knowledgeto see current learning status
Source
git clone https://github.com/aiskillstore/marketplace/blob/main/skills/0xrdan/learn-on/SKILL.mdView on GitHub Overview
Learn On activates continuous learning mode so the router automatically extracts insights during a session. When active, it monitors query activity and triggers extraction after 10 queries or 30 minutes, appending findings to the knowledge base without manual intervention.
How This Skill Works
When Learn On is enabled, the router watches each query and increments a counter. After the threshold of 10 queries or 30 minutes is reached, it runs the extraction process and saves insights to knowledge/learnings/patterns.md, knowledge/learnings/quirks.md, and knowledge/learnings/decisions.md. The mode persists across the session and updates knowledge/state.json with learning_mode, learning_mode_since, and resets queries_since_extraction.
When to Use It
- During long sessions where you want insights captured automatically without manual /learn calls.
- When you need continuous knowledge-base growth (patterns, quirks, and decisions) across a session.
- In multi-topic conversations where missing insights is costly and auto-extraction helps.
- During onboarding or research chats to surface recurring issues automatically.
- When you want to keep a persistent learning state that survives the session and updates the knowledge base.
Quick Start
- Step 1: Update knowledge/state.json to enable learning_mode: true and set learning_mode_since to a current ISO timestamp; reset queries_since_extraction to 0.
- Step 2: Save the updated state back to knowledge/state.json.
- Step 3: Confirm to the user that Continuous Learning: ENABLED and that automatic extraction will trigger per the thresholds.
Best Practices
- Use Learn On for extended discussions to avoid missing insights.
- Keep thresholds (10 queries, 30 minutes) aligned with your session pace.
- Periodically review the generated files at knowledge/learnings/ for accuracy.
- Remember Learn On persists across the session but resets on new sessions.
- Use /learn-off to disable automatic extraction when you don’t want it.
Example Use Cases
- A support chat auto-captures common issues into patterns.md, quirks.md, and decisions.md.
- A product planning call automatically documents recurring design quirks and decisions.
- A developer debugging session surfaces frequent error patterns without manual prompts.
- Customer success conversations save best practices into the learning docs.
- A multi-topic research chat maintains continuous insight capture across topics.
