Prompt Clarity Engine · tabarc-code bundle Rewrites vague or underspecified prompts into clear, LLM-friendly instructions, using a categorised taxonomy of 200+ instruction verbs (reading, extraction, synthesis, evaluation, verification, risk detection, planning, iteration and more), a fixed instruction order, and a six-question completeness check. Diagnoses catch-all wording such as "ingest", "process", "understand", "improve", "optimise", "fix", "continue" and "research", and reconstructs the prompt around an observable action, explicit scope, a stated output form and a validation step. Use whenever someone wants a clearer prompt for an LLM, asks "how do I ask this better", "make this prompt more specific", "why isn't the model doing what I want", or pastes a rough instruction and wants it tightened or made implementation-ready. Also covers prompt engineering audits and turning a one-line ask into a properly scoped brief.
