product-title-generation

Product Title Generation

Use this skill to turn a product name, feature description, technical capability, platform module, learning service, activity theme, or mixed Chinese-English concept into compact Chinese product titles.

The output is a recommendation plus six alternatives, designed for UI entry names, module names, product cards, page section titles, or brand-like short names.

Ownership

MUST generate short product, feature, module, entry, or brand-short titles only. MUST NOT edit files, create brand strategy, check trademark availability, write long marketing copy, or rename code identifiers. For those adjacent requests, MUST stop title generation and route using the "Do Not Use" mapping.

Do Not Use

Route adjacent requests before generating titles:

• Slogans, taglines, ad copy, landing-page copy, or long marketing copy -> use a copywriting workflow.
• SEO page titles, keyword headlines, or search-snippet optimization -> use an SEO/content workflow.
• Trademark availability, legal clearance, naming conflicts, or registration advice -> tell the user this needs legal review.
• Code identifiers, variable names, package names, class names, or refactor naming -> use a code naming/refactor workflow.
• Full brand strategy, positioning, naming architecture, tone system, or brand book work -> use a brand strategy workflow.

Execution

CREATE A TODO LIST FOR THE TASKS BELOW. Keep the list internal unless the user asks to see process.

1. Read the user's input. If no product, feature, concept, or direction is provided, ask one concise question for the missing target and stop.
2. If the input matches any "Do Not Use" case, state that this skill only generates short product titles, route using that mapping, and stop.
3. If the user's title or naming request is ambiguous between a short product title, SEO headline, slogan, campaign copy, full product name, or brand strategy, ask one concise clarification question and stop.
4. If the input contains multiple unrelated products, split them into separate targets and use one "Multi-Product Output Format" block per clear target; if any target is unclear, ask the user to choose the target and stop.
5. For each clear target, normalize the input into four fields: product object, core capability, use scenario, and desired tone. If a field is missing, infer it from the provided text without inventing unrelated positioning.
6. If the product object exists but core capability and use scenario cannot be inferred from the input without inventing unrelated positioning, ask one concise question for the missing capability or scenario and stop.
7. Route the title style for each target. If product object, source brand, scenario, and core capability imply different routes, prioritize core capability first, then use scenario and desired tone to refine wording:
   • Learning products -> companionship, sprint, rescue, training, improvement.
   • Technical platforms -> base, platform, hub, engine, cockpit, infrastructure.
   • Data or observability products -> observation, insight, monitoring, tracing, visibility.
   • Launch or incubation products -> launch, incubation, startup, publishing, product desk.
   • Code or development products -> repository, code, understanding, navigation, insight.
   • Otherwise -> use a neutral product-entry style.
8. Generate and refine candidates until the final visible set contains exactly one recommendation and six unique alternatives, unless the user explicitly requests a different count. Each candidate MUST preserve the core object or core capability.
9. Filter candidates with the title rules below. Remove titles that are too long, too generic, too marketing-heavy, awkwardly translated, or semantically off-target. If too few valid titles remain for one recommendation plus the required number of alternatives, generate more candidates and repeat filtering until the output can be filled. If the user's explicit constraints make the required count impossible, ask one concise clarification question or state the conflict and stop.
10. Select the recommendation using this priority order: semantic fit, product-entry feel, compactness, distinctiveness, and natural Chinese phrasing.
11. Output the exact format in "Output Format" or "Multi-Product Output Format" and end, unless the user explicitly asks for analysis, more options, fewer options, or a different format.