lore

What This Skill Does

The Lore skill serves as an advanced knowledge management layer for the OpenClaw AI agent. It transforms the agent from a stateless conversationalist into a contextual powerhouse by providing an interface for a persistent, searchable research repository. Lore allows the agent to ingest raw information—such as meeting notes, internal documents, and interview transcripts—and retrieve that data with high-fidelity citations. By prioritizing evidence-based responses, Lore ensures that the agent provides answers grounded in verifiable data rather than general training knowledge.

Installation

To integrate Lore into your OpenClaw environment, use the provided command within your terminal or interface:

clawhub install openclaw/skills/skills/mishkinf/lore

This command retrieves the package directly from the openclaw/skills repository. Ensure your environment has the necessary permissions to pull from this repository before initiating the install.

Use Cases

Lore is designed for environments where information longevity is critical:

1. Organizational Memory: Capturing key decisions made in Slack or Notion to ensure future team members understand the "why" behind historical architectural choices.
2. Synthesis of Complex Projects: Using the research tool to analyze multiple project documents to generate status reports, requirement lists, or performance summaries.
3. Meeting Transparency: Ingesting meeting transcripts immediately after sessions to make them searchable, ensuring action items and commitments are tracked and retrieved later.
4. Support and Customer Feedback: Storing user pain points to perform semantic searches, allowing developers to see trends in user friction across months of communication.

Example Prompts

1. "Check Lore for our team's decision on the database migration last quarter and summarize why we chose PostgreSQL over MongoDB."
2. "I'm pasting our latest sync notes from Google Meet. Please ingest these into the 'Project Apollo' knowledge base."
3. "Research the common frustrations mentioned by users regarding our login flow by looking through our support feedback records in Lore."

Tips & Limitations

• Efficiency: Distinguish between search and research. Only use research when cross-referencing multiple documents is absolutely necessary, as it carries a 10x compute cost increase.
• Maintain Idempotency: Don't worry about duplicating entries; the ingest function is idempotent and will automatically deduplicate identical content.
• Contextual Accuracy: Always use get_source when you need to verify specific phrasing. The search preview might exclude nuances found in the full text.
• Precision: Utilize semantic search for conceptual exploration and keyword search when you need to locate a specific identifier or name.