Example Scenarios and Prompt Library
The prompt library is designed for onboarding, validation, and regression testing of assistant behavior.
Library Structure
Prompt categories mirror in-code examples in her-core/her_telegram/handlers.py (_EXAMPLE_PROMPTS):
- chat
- memory
- scheduling
- automation
- web
- mcp_tools
- sandbox
- admin
- personality
- productivity
Telegram shortcuts:
- /example
- /example all
- /example <topic>
Example Scenarios
1. Conversational continuity
Remember that I prefer concise replies.
What should I focus on this afternoon?
2. Reminder scheduling
Remind me in 45 minutes to send the draft.
3. Rule-based automation
Check BTC every 5 minutes and notify me if it rises 8%.
4. Tool-assisted diagnostics
Check SSL expiry for github.com and summarize result.
5. Admin operations
/schedule list
/mcp
Prompt Quality Guidelines
- Be explicit about timeframe, threshold, and expected output format.
- For workflows, include source URL and clear condition.
- For diagnostics, specify host/domain and desired summary level.