Ecommerce.Google Patent Search

ecommerce.google-patent-search

Use this skill to query Google Patents public search. Read references/api.md before constructing a request.

Request contract

• Endpoint: POST /api/v1/tools/research/googlePatent/search
• Required: q (non-empty query, maximum 1000 characters).
• Optional: num (10-100), page, country, language, before, after, sort, type, status, patents, scholar, litigation, inventor, assignee, clustered, dups.
• Send only documented fields; reject unknown fields before network access.

Minimal example:

jsonCopy code
{"q":"wireless earbuds","num":10,"page":1}

Workflow

1. Confirm the requested public entity, market, operation, filters, and expected cost.
2. Validate parameters against the API reference. Send only documented business fields.
3. Run python scripts/google_patent_search.py '<JSON parameters>' --no-cache once. Do not probe alternate products, markets, pages, or operations after a paid failure without user approval.
4. Preserve the full response file. Distinguish the Nexscope transport envelope from the inner business response and report traceId on errors.
5. Summarize only returned facts. Preserve missing values as unknown and identify the requested market and operation.

Cost

Do not reuse the source Skill's point value. This operation consumes Nexscope credits. Preserve X-Cost-Token and X-Cost-Credit from the response headers as server-reported billing metadata. The exact charge is unknown before the first live response.

Error and credit handling

• HTTP 401 means Nexscope authentication failed. Verify NEXSCOPE_API_KEY and NEXSCOPE_PROXY_BASE; do not ask the user to paste credentials into chat or operation JSON.
• HTTP 402 means the account lacks credits. Stop the workflow and direct the user to the access-help page below.
• Marketplace authorization failures require the platform-specific account or token to be renewed. Do not substitute a different store, region, or creator automatically.
• For ambiguous network failures, report whether the attempted operation was a read or mutation. Never repeat a mutation without reconciling its upstream state.

Response handling

For the research HTTP response, require a numeric outer code equal to 0 before using data. Nonzero codes are platform failures; display outer msg without interpreting its wording as a retry instruction. Nexscope preserves upstream messages and translates Chinese to English; successful responses may have msg: null. HTTP 200 alone and nested business code / status do not replace the platform check. See references/api.md for response paths and task-specific states.

Package scripts that unwrap the response keep the platform code and msg under _nexscope; their remaining fields are the business payload. _nexscope.cost is elapsed time, not credits.

Authentication

Set the NEXSCOPE_API_KEY environment variable. If credentials are missing or expire, visit https://www.nexscope.ai/help/skills-external-access?co-from=skillNS to top up credits.

Presenting results

• Lead with the requested entity, market, operation, and the most decision-relevant returned fields.
• Preserve source currencies, units, identifiers, dates, and missing values; do not invent conversions or defaults.
• Keep the full JSON artifact and present compact tables for repeated records when useful.
• Report the X-Cost-Token, calculated Nexscope credits, and trace ID from the saved billing metadata.

Boundaries and privacy

• This migration is public-data and read-only. Do not invoke seller-console, store-authorized, favorite-write, or account mutation operations.
• One explicit request per call. Do not silently change entity, market, date, operation, or page after an error.
• Redact credentials, signed URLs, cookies, and internal account identifiers from user-facing output and saved request examples.
• Do not claim live, ZIP-install, or production validation unless the matching evidence exists in references/testing.md.

References

• Read references/api.md for the complete parameter, response, and error contract.
• Read references/testing.md before executing a live request or claiming a validation level.