PCW
Prompt & Context Workbench
Structure instructions, control context, estimate token use, and define evaluation cases · v1.0.0
Private drafting mode: This workbench does not call an LLM or transmit prompt text. Token counts and quality checks are heuristics, not model guarantees. Do not paste secrets, credentials, regulated records, or data you are not authorized to process.
Prompt architecture
Instructions, task, context, and output contract
Evaluation set
Representative test cases
Version history