discord-soul

What This Skill Does

The Discord Soul skill transforms your static Discord server history into a dynamic, intelligent agent. By ingesting your community's conversation history, the agent learns the unique lexicon, inside jokes, and personality traits of your members. It is not just a chatbot; it acts as an extension of your community's identity. Using a robust multi-stage security pipeline, the skill ensures that all ingested data is sanitized, protecting the agent from prompt injections, jailbreaks, and malicious user behavior found in public logs. Once deployed, the agent can answer historical queries, summarize community sentiment, and engage in conversations that feel authentic to your group.

Installation

To begin, ensure you have the OpenClaw agent installed and access to your Discord server's history. Run the following command in your terminal:

clawhub install openclaw/skills/skills/kgeesawor/discord-soul

Next, you must use the DiscordChatExporter to pull your messages into a local SQLite database. Once exported, use the provided security pipeline script ./scripts/secure-pipeline.sh to sanitize the logs against regex patterns and Claude Haiku-based semantic evaluation. Finally, run the agent creation script specifying your Guild ID to initialize your agent's memory bank.

Use Cases

• Community Archiving: Preserve the legacy of your server by allowing the agent to answer questions about past events or decisions.
• Onboarding Assistance: Automate the "FAQs" of your server; let the agent explain the "who, what, and why" to new members.
• Sentiment Analysis: Ask your agent to reflect on the community's general attitude toward recent server changes or events.
• Creative Writing: Generate content, memes, or updates that align perfectly with your community's established voice and humor.

Example Prompts

1. "What was the general consensus in the #announcements channel regarding the new server roles last month?"
2. "Explain the origin story of our community inside joke about the 'midnight pancake disaster'."
3. "Summarize the historical evolution of our community's coding standards since the server started."

Tips & Limitations

Tips: Regularly run the update script as a cron job to ensure the agent stays current with daily conversations. Always prioritize the semantic evaluation layer in the security pipeline to minimize the risk of malicious adversarial inputs.

Limitations: The agent is only as good as the exported data. Private channels or deleted messages will not be available. Additionally, the agent's performance depends on the density of the training data; very small servers may require more "conversation time" before the agent develops a distinct personality.