--- name: choose-a-domain description: Interview the user about naming strategy, generate product or project names, check domain availability and relevant public conflicts, and rank a shortlist. Use when the user wants help choosing or validating a name and domain. --- # Choose a name and domain ## Interview Use the user's existing brief. Ask focused questions about anything that could change which names qualify or rank highly. - What is being named, who is it for, and what should the name tell them? What should it still fit if the project expands? - What naming strategy fits the brand? Ask about descriptive versus evocative names, favorite and rejected examples, meaningful words or themes, how many word units should be apparent, and any length or spelling preferences. - Which audiences, languages, or markets matter? What existing brands, products, and adjacent categories should inform the conflict check? - Which extensions, registrars, budgets, renewal costs, and premium or aftermarket options are acceptable? Are matching public handles important? - What does a successful shortlist look like? Ask which qualities matter most, what evidence the user expects, and how many names are useful. Confirm the key requirements and separate firm constraints from preferences. If the user asks to approve the brief before research, wait for their answer. ## Generate and verify 1. Generate names across the agreed strategies. Apply firm constraints first. Keep enough varied candidates to learn which patterns work, and preserve the reasons for exclusions. 2. Check the exact domains with the agreed registrar. Record the result URL, time, registration price and term, renewal price, and any disclosed fees or premium status. An RDAP result can help triage registration status. A domain qualifies as available only when the registrar shows a registration offer that meets the brief. Label blocked or ambiguous checks unverified. Keep the lookup read only. 3. Search strong candidates as an exact compact string, as separated words, and with terms from the relevant field. Look for exact and confusingly similar names across domains and relevant products, companies, repositories, and communities. Follow plausible matches to primary pages. Record generic or unrelated search noise separately from conflicts. Disclose any search engine fallback or coverage gap. If a service presents a challenge, stop there and use an agreed alternative when possible. 4. Check public handles only if requested or relevant to the brief. Report what the public page shows. Missing profiles can have reserved usernames. 5. Recheck shortlisted domains if their availability evidence is stale or inconsistent. Public conflict research can inform selection. Formal trademark clearance requires a separate process. ## Rank and present Agree on the scoring priorities before judging. When a numerical comparison helps, score memorability, ease of saying, and ease of spelling and finding. Use independent judges for the lenses when available and useful. Give each the same brief and finalist pool, with search and registrar evidence withheld. If balance matters, use the sum of the three 1–10 scores minus their spread. Disclose any different rubric or scoring limitation. Give the user a concise ranked shortlist with the strongest case and the tradeoff for each leading name. Include the domain, observed registration and renewal prices, check time, relevant conflicts or noise, and the basis for the ranking. Separate unverified names from the available shortlist and summarize why strong candidates were ruled out. Preserve a decision log when the candidate pool spans several rounds.