A production-grade AI learning assistant that facilitates a plan-learn-practice-assess loop with proactive automation and deep subject isolation.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install learning-coach
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install learning-coach using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Learning Coach?
The Learning Coach is a sophisticated educational framework designed for Openclaw Skills that transforms an AI agent into a dedicated tutor. It manages multiple subjects independently, ensuring that progress metrics, quiz histories, and study plans remain isolated for focused growth. By moving away from static materials, it uses Large Language Models to generate dynamic quizzes and provide high-quality, rubric-based grading that offers genuine coaching insights rather than just scores.
This skill is built for users who require structured learning guidance over time. It leverages proactive automation through cron jobs to keep learners on track with reminders and weekly summaries. By integrating content discovery from YouTube and web feeds, it ensures that your study materials are always fresh and relevant to your specific goals and proficiency level.
Learning Coach Use Cases
- Creating a structured learning roadmap for complex technical subjects or professional certifications.
- Automating the discovery and curation of high-quality educational resources from YouTube and RSS feeds.
- Generating challenging, context-aware quizzes to test knowledge retention across different Bloom's levels.
- Receiving detailed, rubric-based feedback on open-ended answers to identify and correct weak concepts.
- Maintaining a consistent study habit through proactive reminders and automated weekly progress reporting.
How Learning Coach Works
- The user defines a subject and goals, which the skill initializes in a dedicated data directory to maintain subject isolation.
- The skill generates a structured weekly plan and daily tasks based on the user's level and time budget.
- Content discovery scripts ingest and rank external resources from YouTube and the web to curate a custom reading/watching list.
- Upon request, an LLM generates a unique quiz containing metadata, rubrics, and feedback rules for the specific topic.
- The learner submits answers, which are then graded by the LLM using the predefined rubric to generate corrective drills.
- An intervention policy analyzes performance metrics to recommend speeding up, stabilizing, or slowing down the learning pace.
Learning Coach Setup
To deploy this skill within Openclaw Skills, follow these initialization steps:
- Run the bootstrap script to ensure all dependencies are met:
python scripts/bootstrap.py
- Create a new subject directory and profile:
python scripts/subject_store.py --create <subject-slug>
- (Optional) Set up proactive reminders by providing cron consent:
python scripts/setup_cron.py --apply
Learning Coach Data Schema & Taxonomy
The Learning Coach maintains a strict subject-segregation model. All learner state is stored under data/subjects/<subject-slug>/ with the following schema:
| File | Description |
|---|---|
profile.json |
Stores learner goals, current level, and weekly time budget. |
plan.json |
Contains the active weekly plan and specific daily tasks. |
quiz-history.json |
A record of LLM-generated quizzes, rubrics, and user attempts. |
progress.json |
Tracks rolling metrics, confidence trends, and identified weak concepts. |
curation.json |
Lists recommended links with rationales and estimated consumption time. |
coach-config.json |
Global settings for cadence and preferred output styles. |
Learning Coach Advanced Features
- Proactive automation via idempotent cron management for reminders and autonomous status updates.
- Adaptive intervention rules that utilize Exponential Moving Averages (EMA) to adjust learning intensity.
- Multi-source ingestion engine that normalizes feeds from YouTube, X (formerly Twitter), and web RSS into candidate JSON.
- Bloom's Taxonomy-aware assessment generation to ensure depth of testing beyond simple multiple-choice questions.
- Explicit consent-based workflow for autonomous agent actions to ensure transparency and user control.
SKILL.md
Loading
Related Openclaw Skills
Intercom
A secure, opt-in communication bridge for routing messages between different AI agent sessions via internal session keys.
Xiaomi Outbound Bot (Alibaba Cloud)
A specialized automation skill for triggering and managing batch outbound telephone tasks via the Alibaba Cloud Xiaomi platform.
Business Claw ERPNext Skills
A powerful collection of pre-defined business workflows that integrate MCP tools with ERPNext to automate complex enterprise tasks.
Sunday
Sunday provides AI agents with a dedicated email address and an end-to-end encrypted credential vault for fully autonomous identity management.
OpenBio
A comprehensive API and toolset for biological data access, literature search, and advanced computational biology workflows.
hey.lol AI Creator Skill
Enable your AI agent to build a social media presence, engage with humans, and earn money through paywalled content.
Featured*
