Prompt Safety Notes
#### Prompt Safety Notes
The prompt examples focus on controlled classification, structured extraction, and review safeguards.
##### Use Boundaries
- Prompt results should be treated as draft review support.
- Prompt results should not be used as final decisions.
- Missing or unclear information should be flagged for analyst review.
- Records should not be closed, reprioritized, or merged based only on prompt output.
- Final recommendations should remain traceable to the source record and validation notes.
##### Review Expectations
Prompt-assisted review should help organize information, not replace judgment. Classification, prioritization, and recommendations should remain subject to manual review.when to use it
Community prompt sourced from the open-source GitHub repo Tyler-Windes/ai-assisted-workflow-analysis-demo (MIT). A "Prompt Safety Notes" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.
tags
roleplaycommunitygeneral
source
Tyler-Windes/ai-assisted-workflow-analysis-demo · MIT