brand-name-clearance

name	brand-name-clearance
description	Evaluate candidate names for projects, products, companies, open-source repos, apps, or SaaS brands. Use when the user asks to name or rename something, compare brand names, check trademark/domain/social/package conflicts, assess brand risk, or produce a naming clearance and marketing report.

Brand Name Clearance

Overview

Use this skill to turn candidate names into a practical clearance report: trademark signals, domain availability, developer/package namespace occupancy, web duplicate risk, brand confusion risk, and launch/marketing guidance.

This is a screening workflow, not legal advice. Always state the check date, cite web sources used, and say that final clearance should be confirmed through official trademark databases, registrar checkout pages, and counsel when the project is commercial or high-stakes.

Quick Start

Run the bundled checker first, then supplement with live web search:

python3 scripts/brand_check.py \
  --names "NameOne" "NameTwo" \
  --tlds com,ai,io,dev,app \
  --output markdown

If the user has no names yet, generate 10-20 candidates, eliminate obvious collisions by live search, then run the script on the survivors.

Workflow

1. Scope the naming problem

   • Identify what is being named: repo, app, company, open-source project, SaaS, CLI, package, mobile app, or umbrella brand.
   • Note target markets when known: US, China, EU, global, developer tooling, video, AI, consumer SaaS, etc.
   • Ask only if essential information is missing. Otherwise assume global web/software screening, with .com and one category TLD such as .ai prioritized.

2. Run deterministic checks

   • Use scripts/brand_check.py for domains, Markbase USPTO-index screening, GitHub, npm, PyPI, crates.io, Docker Hub, Hugging Face, and weak social handle probes.
   • Treat script output as an initial signal. HTTP blocks, WHOIS differences, and registry delays can produce unknown or stale results.

3. Run live web search

   • Use exact quoted searches for each candidate:
     • "Name"
     • "Name" app
     • "Name" software
     • "Name" AI
     • "Name" video editor or the user's category
     • "Name" trademark
   • Search confusingly similar names when the candidate has obvious neighbors, such as missing vowels, swapped letters, same pronunciation, or a famous adjacent brand.
   • Prefer primary sources: official product sites, trademark databases, package registries, GitHub, app stores, social profiles, and domain/registry pages.
   • See references/search-playbook.md for query patterns and platform URLs.

4. Assess trademark and confusion risk

   • Highest risk: exact or near-exact mark in software, SaaS, AI, video, media, developer tools, or the user's category.
   • Medium risk: same pronunciation, one-letter difference, similar spelling, same field, occupied .com, or occupied main social handles.
   • Lower risk: unrelated personal names, old events, OCR noise, distant industries, or generic dictionary mentions.
   • Check Nice classes most often relevant to software: 9 (software/downloadable apps), 42 (SaaS/software services), 41 (media/video/education/content), plus any user-specific class.

5. Assess brand and marketing risk

   • Score memorability, pronunciation, spelling, SEO uniqueness, category fit, extensibility, and whether the name can become a commercial brand.
   • Penalize names that are too descriptive, too close to upstream projects, hard to hear over voice, hard to spell from memory, or already noisy in search.
   • Recommend a launch posture: open-source project name, commercial product name, GitHub org/package name, domain bundle, and social handle strategy.

6. Write the report

   • Use the structure in references/report-template.md.
   • Lead with the practical recommendation, then show evidence.
   • Include a ranked shortlist and an avoid list.
   • For each recommended name, include next actions: domains to buy, handles to claim, trademark databases to verify, and marketing angle.

Output Rules

• Be decisive, but never say a name is legally safe.
• Distinguish available, taken, unknown, and needs manual verification.
• Explain why a collision matters: same category, same pronunciation, SEO noise, upstream confusion, social handle loss, or weak commercial posture.
• If using web search, include links to sources in the final answer.
• If no candidate is good, say so and propose new directions instead of forcing a weak recommendation.

Bundled Resources

• scripts/brand_check.py: repeatable initial checker for domains, package registries, selected platforms, and Markbase trademark screening.
• references/search-playbook.md: live search patterns, source priorities, and manual verification links.
• references/report-template.md: recommended structure for final reports.