ROLE: You are a research engineer building a reproducible decision tool.
CONTEXT: Fictional data: plan A is $12 monthly for 100 credits; B is $20 for 200 credits. Required usage is 150 monthly. Credits expire; overages are unavailable. These are fixed practice inputs, not live prices.
TASK / OBJECTIVE: Compare feasible options and build a semantic HTML decision table using locally compiled Tailwind and vanilla JavaScript to change usage.
CONSTRAINTS: Use supplied facts only. Mark infeasible plans; do not invent overage fees. Do not browse or add a backend. Validate non-negative integer usage. Keep original inputs visible.
FORMAT / OUTPUT SPEC: Return assumptions, comparison, complete files, build commands, and tests. Separate calculations from commentary and executed checks from untested work.
EXAMPLES: At 150 credits, only B is feasible, at $20. At 201, neither qualifies. Blank or negative input gets a validation message, not a price.
ITERATION & EVALUATION: Check 0, 100, 101, 200, and 201 credits, keyboard use, and a 360 px viewport. Do not describe this demo as current market research.

Source: https://makersignal.net/guides/ai-prompting/#code-research-template