| name | naming-studio-research |
| description | Research external facts for a Naming Studio request. Use when a naming plan step references naming-studio-research, or when candidate names must avoid currently popular Chinese baby names or be audited for negative Mandarin homophones. |
Naming Studio Research
This skill gathers facts that Codex cannot know reliably without looking: current name popularity and real-world homophone associations. It is always invoked as a step inside a naming request; it never writes to the GUI itself — naming-studio-generate owns save_naming_product_result.
Topic: popular-names
Triggered by the GUI filter 避开热门名字.
- Web-search recent rankings of popular Chinese newborn names (e.g. 公安部全国姓名报告、近三年新生儿爆款名字榜单).
- Compile a blocklist of the top given names (both 双字 and 单字 forms).
- Return the blocklist to the calling flow. Candidate names must not match any entry, and near-misses (same characters reordered, same pinyin with identical tones) should be flagged.
Topic: mandarin-homophone
Triggered by the GUI filter 避开普通话谐音.
- For each candidate full name (surname + given), enumerate its pinyin and tone sequence.
- Check for negative homophones, mocking nicknames, brand or slang collisions. Search the web when uncertain — playground nicknames evolve faster than training data.
- Reject or replace any candidate with a plausible negative association; note the reason.
Constraints
Findings feed the generation step of the same request. Keep research concise: one search pass per topic, summarize into actionable lists, no long reports in chat. Do not ask for API keys; use built-in web search capability only.
