tools-only CLEAR Framework
The CLEAR framework evaluates and improves prompts and PRDs. CLEAR was developed by Dr. Leo Lo and published in the Journal of Academic Librarianship (July 2023). The acronym stands for Concise, Logical, Explicit, Adaptive, and Reflective.
Components
| Component | Focus | Typical improvements |
|---|---|---|
| Concise | Reduce noise and pleasantries | Remove filler, tighten language, emphasize action verbs |
| Logical | Improve flow and ordering | Restructure prompts into context → requirements → constraints → outputs |
| Explicit | Clarify expectations | Specify persona, tone, output format, success criteria, examples |
| Adaptive | Offer alternative approaches | Provide variations, alternative structures, temperature suggestions |
| Reflective | Encourage validation | Add checklists, edge cases, fact-checking steps, risk mitigation |
Applying CLEAR
When evaluating a prompt or PRD, work through each component:
- Concise: Score Concise, Logical, and Explicit — produce a single improved prompt and a list of labeled changes.
- Deep: Unlock Adaptive and Reflective components to deliver alternative phrasings, structures, and validation checklists.
- PRD review: Validate generated quick PRDs for the C/L/E components unless validation is explicitly skipped.
- Summarize: Optionally re-run CLEAR on extracted prompts to give an optimized variant ready for AI agents.
Further reading
- Framework guide: https://guides.library.tamucc.edu/prompt-engineering/clear
- Research paper (PDF): https://digitalrepository.unm.edu/cgi/viewcontent.cgi?article=1214&context=ulls_fsp