# elizaOS Development Discord - 2025-05-22

## Overall Discussion Highlights

### Agent Development & Troubleshooting
- A Bitcoin-focused agent was posting "\n\n" in tweets instead of proper line breaks
- The issue was traced to an XML template explicitly instructing the agent to "Use \n\n (double spaces) between statements"
- GitHub API rate limiting issues were encountered when trying to update the plugin registry

### Documentation Needs
- Request for documentation about "llms-full.txt" file and its role in the ecosystem
- Ruby explained it functions as a "brain-menu" for characters to select which language model to use for different cognitive tasks
- Documentation for custom plugin development was requested but remained unanswered

### Resource Sharing
- Multiple visual briefings and knowledge posters were shared by "jintern"
- Resources included "daily-silk.png", "ainews-elizaos-discord.png", and "hackmd-facts-briefing.png"
- A request was made for the AutoFun brand kit that had been previously shared

### Pull Requests
- Two PRs (#4706 and #4707) were shared to the elizaOS/eliza repository
- Received skeptical responses without substantive technical discussion

## Key Questions & Answers

1. **Q: Why is my agent posting "\n\n" instead of a line break?**  
   **A:** That usually means there's an extra newline character in your response template or the text processing isn't trimming whitespace properly. The issue was identified in the XML template instructions that explicitly tell the agent to "Use \n\n (double spaces) between statements."

2. **Q: What is llms-full.txt and how does it operate in the ecosystem?**  
   **A:** It's a "brain-menu" for characters, letting them pick which LLM to use for different cognitive tasks. It interfaces with the knowledge system and memory management.

3. **Q: Does eliza still run `BAAI/bge-small-en-v1.5` for local embeddings?**  
   **A:** User found the answer independently without community assistance.

## Community Help & Collaboration

1. **Agent Tweet Formatting Issue**
   - **Helper:** Ruby
   - **Helpee:** bitcryptowski.btc
   - **Context:** Agent posting "\n\n" instead of line breaks in tweets
   - **Resolution:** Identified the issue in the XML template instructions and suggested modifying the template to change '\n\n' to '\n' or remove newlines entirely

2. **GitHub API Rate Limiting**
   - **Helper:** Ruby
   - **Helpee:** bitcryptowski.btc
   - **Context:** GitHub API rate limit errors when updating plugin registry
   - **Resolution:** Suggested creating a GitHub personal access token and exporting it as GITHUB_TOKEN environment variable

## Action Items

### Technical
- **Modify XML response template:** Change '\n\n' to '\n' or remove newlines entirely for proper tweet formatting (Mentioned by Ruby)
- **Create GitHub personal access token:** Export it as GITHUB_TOKEN to resolve rate limiting (Mentioned by Ruby)
- **Clear plugin cache:** Run 'rm -rf ~/.elizaos/plugins' before updating plugins (Mentioned by Ruby)

### Documentation
- **Create documentation explaining llms-full.txt:** Detail its role in the ecosystem (Mentioned by Hidden Forces)
- **Create documentation for custom plugin development:** Provide guidance on how to create custom plugins (Mentioned by Not Eric)
- **Make AutoFun brand kit easily accessible:** User mentioned it was shared before but couldn't locate it (Mentioned by ElizaBAO🌟)

### Feature
- **Implement Twitter profile search functionality:** Need ability to search and access Twitter user profiles directly (Mentioned by buttercup)