mcpField
A
amazon-research

amazon-research

github.com/postplusaiMarketing & growth27319 installsUpdated 1d ago
Repository

About amazon-research

An instruction skill for collecting and analyzing Amazon product, ASIN, review, and bestseller evidence. It guides bounded research workflows and requires results to identify their Amazon evidence sources.

Key features

  1. 01Classify product, ASIN, review, and bestseller research
  2. 02Compile requests for hosted collections
  3. 03Normalize and analyze collected data
  4. 04Keep initial research scope bounded
  5. 05Report evidence sources and missing data

Use cases

  1. 01Compare Amazon marketplace prices and sellers
  2. 02Analyze reviews for recurring complaints
  3. 03Map products on bestseller lists
  4. 04Enrich a known list of ASINs

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/postplusai/postplus-skills --skill amazon-research

After installing, type/amazon-researchin Claude Code to confirm it loaded.

Details

From the project README

Amazon Research

Use this skill only when the request explicitly needs Amazon platform evidence: ASINs, reviews, best sellers, seller listings, marketplace pricing, or Amazon competitor data.

Apply shared rulebook and user-guidance rules from postplus-shared.

Do Not Use When

Generic category, market, or positioning questions should start with cheaper general research or strategy skills unless the user names Amazon, provides Amazon URLs or ASINs, or asks for marketplace price/review evidence.

Task Shapes And Keys

Classify the request first:

  1. Product discovery
    • keywords, category pages, search URLs
    • collection key: amazon-products
    • alternate key: amazon-free-products
  2. ASIN enrichment
    • known ASIN list, title/price/rating/seller details
    • collection key: amazon-asins
  3. Review mining
    • reviews, complaints, low-star patterns
    • collection keys: amazon-reviews, amazon-reviews-v2
  4. Bestseller mapping
    • trend products or leaderboard products
    • collection key: amazon-bestsellers

Hosted collection input must be a schemaVersion: 1 envelope whose input field contains the compiled Amazon request.

Default Workflow

Use the lightest valid chain:

  1. compile a small request from keywords, URLs, or ASINs,
  2. normalize the dataset,
  3. analyze ranking, pricing, reviews, or seller signals,
  4. synthesize the result and name missing evidence.

Start with one keyword set, one ASIN batch, one bestseller page, or at most 20 product-discovery items per start URL. Expand only after the first pass proves useful.

Keep request files, raw datasets, normalized datasets, and caches under .postplus/amazon/; keep only final summaries or shortlisted exports outside .postplus/.

Good Output

Return ranking tables, detail tables, review summaries, price bands, negative review patterns, strongest product examples, and the exact Amazon evidence source behind each conclusion.

Failure Modes

  • Stop if the user has not asked for Amazon evidence and no Amazon identifiers were provided.
  • Stop on unsupported keys, missing auth, unavailable hosted service, stable network failure, or malformed collection output.
  • Do not switch to ad hoc scraping, private backend calls, or generic web articles for Amazon platform-data collection.

Handoff

After delivering results, hand off to sourcing-selection for supply-side validation or continue deeper review mining for a specific ASIN.

Public Command Boundary

  • Choose the smallest matching command or workflow from the user input and run it directly.
  • Readiness diagnostics: postplus doctor --skill amazon-research.
  • If an owned CLI or script command fails, report the exact error and stop. Do not bypass the failure with metadata-only answers, readiness probing, local payload rewrites, fallback providers, or unpublished tools.
  • Use postplus research schema --collection-key amazon-asins --json only when constructing or repairing an unknown request shape.
  • Hosted collection: postplus research collect --skill amazon-research --collection-key amazon-asins --input <hosted-envelope.json> --output <collection-result.json>.
  • Resume a pending collection: postplus research collect --run-handle <runHandle> --output <collection-result.json>.
  • Keep the first pass bounded; expand only after inspecting the first result.
  • If the CLI returns a quote-confirmation challenge, run postplus quote confirm --json --challenge-file <challenge.json> and retry with the returned token.

Comments

Markdown supported