openclaw-knowledge-coach

OpenClaw Knowledge Coach

An OpenClaw-native skill for local knowledge retention. Import knowledge, generate practice, evaluate answers, and produce insight cards — all powered by the host agent's model, with zero extra API key configuration.

OpenClaw Users (Recommended)

When running inside an OpenClaw agent, the host provides model configuration. No praxis llm setup or API key setup is required.

Install the library:

pip install openpraxis

The skill uses the host agent's LLM capability automatically. Set the environment variable to enable OpenClaw mode:

export OPENPRAXIS_MODE=openclaw

Standalone CLI (Fallback)

For use outside of an OpenClaw agent, configure your own provider:

pip install openpraxis
praxis llm setup
praxis llm show

Environment variables override config file values:

export OPENAI_API_KEY="your_key_here"
# or ARK_API_KEY / MOONSHOT_API_KEY / DEEPSEEK_API_KEY based on provider

Core Workflow

1. Confirm scope and source

• Confirm knowledge domains, source folders, and accepted file types.
• Confirm whether to preserve existing metadata (tags, dates, project names).

1. Define import contract

• Normalize each source into a record with doc_id, title, source_path, tags, created_at, and content.
• Split long content into chunks with stable IDs such as doc_id#chunk-001.

1. Import into OpenClaw

• Ingest normalized records into the local OpenClaw knowledge base.
• Keep a deterministic mapping between source file and imported IDs for later updates.

1. Generate exercises at import time

• For each chunk, create at least one retrieval exercise.
• Prefer three exercise types:
  • free-recall: ask the user to explain from memory.
  • qa: ask direct question-answer pairs.
  • application: ask scenario-based transfer questions.
• Save answer keys and concise grading rubrics.

1. Build review queue

• Group exercises by topic and difficulty.
• Schedule spaced review windows (for example: day 1, day 3, day 7, day 14).