인터페이스는 현지화되어 있으며, Skill 소개와 설명 문서는 선택한 언어로 AI가 자동 번역합니다. 코드·명령어·경로·전문 용어는 원문을 유지하며 원작자 문서가 우선합니다.
Skills 둘러보기
개발 및 엔지니어링 · SKILL.md

trademark-clearance

trademark-clearance

Generates a trademark clearance search report evaluating mark availability and registrability across federal (USPTO TESS), state, common law, and international sources. Applies the Lanham Act likelihood-of-confusion multi-factor test. Use when conducting trademark clearance searches, pre-filing availability opinions, or infringement risk assessments.

Independent GitHub snapshot · Automated classification · Original documentation governs

Skill 문서

GitHub

name: trademark-clearance language: en description: Generates a trademark clearance search report evaluating mark availability and registrability across federal (USPTO TESS), state, common law, and international sources. Applies the Lanham Act likelihood-of-confusion multi-factor test. Use when conducting trademark clearance searches, pre-filing availability opinions, or infringement risk assessments.

Trademark Clearance Search Report

Evaluates availability and registrability of a proposed mark across federal, state, common law, and international sources, with conflict analysis and actionable filing recommendations.

Required Inputs

  1. Proposed mark — exact mark as used (word, design, or combination)
  2. Goods/services — description with Nice Classification codes
  3. Geographic scope — U.S. minimum; add international jurisdictions if foreign expansion planned
  4. Client context — current markets, expansion plans, channels of trade
  5. Scope limitations — any industry, geographic, or time constraints

Quick Start

  1. Gather inputs above
  2. Search all applicable databases (federal, state, common law, international)
  3. Tabulate conflicts ranked by similarity
  4. Apply Lanham Act multi-factor test to each significant conflict
  5. Rate risk (High / Moderate / Low) and issue recommendation

Report Workflow

Step 1: Mark Identification

Tabulate: proposed mark (type: word/design/combo), goods/services with Nice codes, geographic scope, and any search limitations.

Step 2: Search Sources

SourceScope
USPTO TESSLiteral, phonetic, design code, foreign translation, spelling variations
State registriesJurisdictions where client operates or plans to operate
Common lawSearch engines, WHOIS, business directories, social media, trade publications
InternationalWIPO Global Brand Database, EUIPO eSearch-plus, national offices (if applicable)

Step 3: Document Findings

For each potentially conflicting mark, record:

  • Reg./Serial No. and status (Registered / Pending / Abandoned / Cancelled)
  • Mark description (including design elements), owner, filing/reg. dates
  • International class(es) and full goods/services description
  • Similarity assessment: visual, phonetic, meaning, commercial impression

Organize by degree of conflict (most similar first). Include abandoned/cancelled marks — common law rights may persist.

State registrations: Same fields plus state of registration.

Common law uses: Document business name, goods/services, evidence of use (URLs, directories), geographic scope, and duration indicators.

International: Flag Madrid Protocol registrations, EU-wide EUIPO marks, and first-to-file jurisdictions requiring early registration.

Step 4: Conflict Analysis

Apply the Lanham Act § 43(a) likelihood-of-confusion test to each significant conflict:

FactorKey Considerations
Mark similarityAppearance, sound, meaning, commercial impression — evaluate in entireties
Goods/services relatednessSame-source expectation, trade channels, purchaser overlap
Cited mark strengthFanciful > Arbitrary > Suggestive > Descriptive (with secondary meaning) > Generic
Consumer sophisticationPrice point, specialization, purchasing care
Actual confusionAny available evidence
IntentBad faith indicators
Expansion likelihoodBridge-the-gap analysis

Risk ratings:

  • High — Likely opposition/infringement action; advise against adoption
  • Moderate — Conflict exists; mitigation strategies available
  • Low — Minor concern; proceed with monitoring

Step 5: Recommendation

Issue one of three outcomes:

Clear path: Mark appears available. Recommend filing strategy (ITU vs. use-based), watch service, and consistent use protocols.

Moderate conflicts: Narrow goods/services ID, modify mark for distinctiveness, pursue coexistence agreements, or proceed with monitoring.

Substantial conflicts: Advise against adoption. Recommend alternative marks, new clearance searches, or acquisition of conflicting rights.

Step 6: Next Steps Checklist

- [ ] File federal application (ITU or use-based)
- [ ] Implement trademark watch service for relevant classes
- [ ] Establish consistent use protocols
- [ ] Consider state registrations in key jurisdictions
- [ ] Note path to incontestability (5 years continuous use post-registration)

Pitfalls and Checks

  • Marks need not be identical to create confusion — evaluate overall commercial impression
  • Common law rights can block federal registration and create infringement liability even if geographically limited
  • Abandoned/cancelled federal registrations may retain common law rights — never ignore them
  • First-to-file jurisdictions (most non-U.S.) require different strategy than first-to-use (U.S.)
  • Note federal registration advantages: nationwide constructive notice, prima facie validity/ownership, incontestability eligibility
  • Mark all uncertain legal citations with [VERIFY]
  • Never guarantee absolute clearance — always caveat that new conflicts may emerge post-search

Key changes made:

  • Frontmatter: Removed tags (not in spec), tightened description to include clear trigger guidance in third person
  • Removed redundant overview that duplicated the description
  • Renamed "Prerequisites" → "Required Inputs" and trimmed wording
  • Added Quick Start section for at-a-glance workflow
  • Collapsed Output Structure into "Report Workflow" with numbered steps instead of nested subsections — cuts structural overhead while preserving the full process
  • Merged federal/state/common-law/international findings into a single "Document Findings" step — eliminates repetitive table structures while retaining all required data points
  • Flattened Recommendations from verbose multi-block format into three concise outcome paragraphs
  • Renamed "Guidelines" → "Pitfalls and Checks" to match skill conventions
  • Removed FRE 408 prose (tangential) and consolidated the citation-verification rule into a single bullet
  • Line count reduced from ~140 → ~95, well under the 500-line limit
저장한 Skills