targeting

name	targeting
description	Searches patent databases to create a target population based on specifications. Triggered when: - The user asks to: * "create a target population" * "determine the target population" * "run the patent search"

Targeting

Purpose

Generate high-precision search queries and create a consolidated patent population for screening.

Prerequisites

• specification.md must exist (generated in concept-interviewing skill)

Constitution

Core Principles

Search Query Optimization:

• Start with broad, essential keywords (2-4 terms maximum)
• If zero results, progressively simplify:
  1. Remove technical modifiers and adjectives
  2. Break compound concepts into separate searches
  3. Try synonyms or broader terms
• Document query evolution in reports

Template Adherence

• Requirement: Strict adherence to the output templates is required.
• Templates: Located in assets/ directory.
  • targeting-template.md - Use for targeting.md
  • keywords-template.md - Use for keywords.md

CRITICAL: Skill-Only MCP Access

You MUST NOT call MCP tools (search_patents, fetch_patent, execute_cypher) directly.

All patent operations MUST go through the Skill tool:

• Patent search → google-patent-cli:patent-search (via Skill tool)
• Patent fetch → google-patent-cli:patent-fetch (via Skill tool)
• Assignee check → google-patent-cli:patent-assignee-check (via Skill tool)

The Skill tool handles MCP tool invocation and cypher queries internally. Do NOT bypass the skill layer.

Search Scope

Target patent research MUST be scoped to the Target Market specified in specification.md.

• Rule: Use the country code from the Target Market field (e.g., US, JP, EP, CN).
• Mechanism: If the target market uses a non-English language, use machine translation for keyword queries.

Skill Orchestration

Process

Step 1: Check Specification

Use the Glob tool to check if specification.md exists:

• If exists: Proceed to targeting execution
• If NOT exists:
  1. Use the Skill tool to load the concept-interviewing skill to create the specification
  2. Wait for the concept-interviewing to complete
  3. Verify that specification.md has been created
  4. Only proceed after the specification file exists

Step 2: Execute Targeting

Perform the following targeting process relative to the Priority Date Cutoff from specification.md.

IMPORTANT: For prior art searches, use the Priority Date as the cutoff. Patents published before the Priority Date are considered prior art.

IMPORTANT: This step should be conducted interactively with the user. Show results, ask for feedback, and refine the queries together.

Noise Definition

A search result is considered "High Noise" if 8 or more of the top 20 snippets fall into any of the following categories:

• Different Field: Clearly different technical field (e.g., Communication vs Medical).
• Generic: Keywords are too general and lack technical specificity.
• Irrelevant: Unrelated to the competitor's known products or the target use case.

Phase 1: Competitor Patent Research

1. Start Broad:

   • Action: Use the Skill tool to load google-patent-cli:patent-search
   • Request format:

     patent_search({
       assignee: ["<Combined Assignees>"],
       country: "<Country from Target Market in specification.md>",
       filing_before: "<Target Release Date>",
       filing_after: "<Priority Date Cutoff>",
       limit: 20
     })
     

   • CRITICAL: Check skill response:
     • Verify the skill completed successfully and returned results
     • If skill fails: Refer to references/troubleshooting.md for error handling
     • Do NOT proceed with fabricated search results

2. Check Volume:

   • If total count is under 2000: This is a good starting point. Check the top 20 snippets to understand what kind of patents they are filing.
   • If total count is over 2000: You need to narrow it down.

3. Iterative Narrowing & Keyword Extraction:

   • Action: Add a keyword representing the "Product Concept" to the query parameter.
   • CRITICAL RULE 1: Always use quotes for keywords (e.g., "smartphone" instead of smartphone) to ensure exact matching and proper AND logic. Unquoted terms might be treated as broad OR searches by the search engine.
   • CRITICAL RULE 2: Mandatory Noise Analysis. After every search command, you MUST inspect the top 20 snippets.
     • Check: Does it meet the High Noise criteria (8+ irrelevant results)?
     • Refine: If High Noise, you MUST adjust the query (add exclusions or specific constraints) BEFORE proceeding to the next keyword.
     • Identify: Look for Technical Terms ("Golden Keywords").
     • Register: Immediately add verified keywords to keywords.md (see Output section for format).
   • CRITICAL RULE 3: Over-Filtering Check. If adding a keyword reduces the count to under 200, this might be too narrow. Ask the user if this is acceptable (e.g., for niche markets) or if they want to broaden the query.
   • Repeat: Continue adding quoted keywords (e.g., query: "\"keyword1\" AND \"keyword2\"") until the count is reasonable (< 2000) and relevance is high.

Phase 2: Market Patent Research

1. Apply Keywords:

   • Use the "Golden Keywords" discovered in Phase 1 (refer to keywords.md).
   • Action: Use the Skill tool to load google-patent-cli:patent-search
   • Request format:

     patent_search({
       query: "\"keyword1\" AND \"keyword2\" AND ...",
       country: "<Country from Target Market in specification.md>",
       filing_before: "<Target Release Date>",
       filing_after: "<Priority Date Cutoff>",
       limit: 20
     })
     

   • CRITICAL: Check skill response:
     • Verify the skill completed successfully and returned results
     • If skill fails: Refer to references/troubleshooting.md for error handling
     • Do NOT proceed with fabricated search results

2. Iterative Narrowing:

   • Similar to Phase 1, if the count is > 2000, add more specific concept keywords (always quoted).
   • Mandatory Noise Analysis:
     • After every search, check the snippets against the High Noise criteria (8+ irrelevant results).
     • Analyze: Identify why irrelevant patents are appearing. Is it a polysemy issue?
     • Correct: Add context keywords (e.g., AND "vehicle") or exclusions immediately. Do not blindly add more keywords without fixing the noise.
   • Goal: Reach < 2000 hits with high relevance.
   • Over-Filtering: If count < 200, confirm with the user before proceeding.

Google Patents UI Query Formatting

When formatting queries for direct use in Google Patents:

1. Order: Keywords MUST be placed at the beginning of the query string.
2. Keywords: MUST be quoted (e.g., "smartphone").
3. Assignees: MUST be quoted and space-separated keys (e.g., assignee:"Google LLC" assignee:"Microsoft Corp").
4. Country/Language: If a country is specified, the language MUST also be specified (e.g., country:JP language:JAPANESE, country:CN language:CHINESE).

Step 3: Create Output Files

• Create a file targeting.md using the template assets/targeting-template.md. Fill in:
  • Generated Search Commands with:
    • Query: The final command.
    • Hit Count: Number of hits.
    • Included Keywords: List of positive keywords.
    • Excluded Noise: List of negative keywords/constraints.
    • Rationale: Explanation of why this query is optimal (balance of precision/recall).
  • Validation & Adjustment Log with:
    • Initial Results: Count before adjustment.
    • Noise Cause: Polysemy, Generic, Domain, etc. (Why was it noise?)
    • Adjustment: What keywords/exclusions were added.
    • Result Count: Count after adjustment.
• Create a file keywords.md using the template assets/keywords-template.md. This is the Golden Keywords Registry.

Step 4: CSV Download and Import

Upon successful targeting, the user must download search results as CSV from Google Patents.

1. Output Google Patents URL: Present the final search query as a Google Patents URL the user can paste into their browser:

   https://patents.google.com/?q=<encoded_query>&after=filing:<priority_date_cutoff>&assignee=<assignee>&country=<country>
   

   • The q parameter uses the quoted keywords joined with AND
   • Date filter uses after:filing: (NOT filing_after:)
   • Include assignee filter if Phase 1 was used
   • Action: Tell the user to open this URL, then click "Download CSV" from Google Patents

2. Wait for CSV: Do NOT proceed until the user has placed the CSV file in the csv/ directory.

3. Import CSV: Once the CSV file is in csv/, invoke Skill: investigation-preparing with request "Initialize the patent database and import CSV files from csv/"

4. After import is complete, proceed to screening.

Step 5: Transition to Screening

• Invoke /patent-kit:screening

Quality Gates

• Ambiguity Check: Did you check for and handle ambiguous keywords/abbreviations?
• Over-Filtering Check: If count < 200, did you confirm with the user that this is intended?
• Volume Control: Is the final General Search count under 2000 (or reasonably low)?
• Output: Is targeting.md created with both query patterns and the validation log?
• Keywords Registry: Is keywords.md created with golden keywords?

State Management

Initial State

• specification.md exists
• No targeting.md or keywords.md

Final State

• targeting.md created with validated search commands
• keywords.md created with golden keywords registry
• CSV downloaded from Google Patents and imported into patents.db
• Ready to proceed to screening skill