Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Amazon Review Gold Miner — Review Intelligence

    1

    Turn Amazon review corpora into evidence-grounded product, listing, quality, and customer insights.

    $9.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Amazon Review Gold Miner — Review Intelligence

    Amazon Review Gold Miner — Review Intelligence

    Example session with this skill installed

    Analyze this explicitly synthetic, 10-review sample for a travel organizer. Five sample reviews mention cables being difficult to find quickly; three praise the compact size; two report a zipper snag; one says the listing led them to expect water resistance, which is not a confirmed product feature. These example theme counts overlap because one review can mention multiple aspects. Show the corpus denominator and caveat that the sample is synthetic, classify aspects and root-cause hypotheses, suggest validation steps and product/listing/support actions, identify product-truth gaps, and state what cannot be concluded. Do not invent quotes, claim representativeness, or treat review statements as verified product facts.

    • Read your context and instructions
    • Compiled the amazon review gold

    Review Intelligence Brief — Synthetic Travel Organizer Sample

    Corpus: 10 explicitly synthetic review examples. This is a demonstration dataset, not real customer feedback or a representative sample. Themes may overlap, so counts do not sum to 10.

    Aspect findings

    ThemeCount / denominatorEvidence classInterpretation
    Hard to locate cables quickly5/10SYNTHETIC-SAMPLECandidate retrieval/friction issue; validate by observed task tests
    Compact size praised3/10SYNTHETIC-SAMPLEPossible benefit worth preserving; check fit and capacity trade-offs
    Zipper snag reported2/10SYNTHETIC-SAMPLEPotential quality signal; inspect product samples, failure conditions, and batch history
    Water-resistance expectation mismatch1/10SYNTHETIC-SAMPLEPossible listing/expectation gap; no confirmed water-resistance feature

    Root-cause hypotheses to test

    • Cable retrieval may relate to compartment layout or visibility; observe users packing and retrieving items before redesign.
    • Zipper snag may relate to hardware, seam alignment, overfilling, or use conditions; inspect returned samples and request supplier QC records. Do not label a confirmed defect from two synthetic reports.
    • The expectation mismatch may come from copy, imagery, customer inference, or another source. Audit all claims and images against the product specification; do not add a resistance claim without defined test evidence.

    Recommended actions

    Product/QC: run repeatable zipper-cycle and load tests on multiple samples and lots; record conditions and failure rate.
    2.

    Usability: test cable retrieval with target users and compare a removable pouch or visible elastic loops as hypotheses. Preserve compactness while measuring capacity trade-offs.
    3.

    Listing/support: review product copy and imagery for implied water resistance; clarify confirmed material and use limits, and prepare a factual support answer.
    4.

    Next review analysis: obtain a complete, dated, variant-matched export; deduplicate, inspect sampling, compare cohorts, and retain exact denominators.

    Confidence: INSUFFICIENT for real-world prevalence or root cause. These illustrative counts cannot establish defect frequency, market demand, causation, return risk, or a business outcome. No customer quote or product fact is asserted.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Amazon Review Gold Miner transforms authorized customer reviews and feedback into decision-grade product intelligence, not just a sentiment summary. It validates marketplace, product and variant identity, language, dates, source, corpus size, sampling method, duplicates, missing text, and truncation before analysis. It classifies sentiment by aspect and extracts customer jobs, desired outcomes, praise, complaints, defects, friction, expectation gaps, compatibility issues, packaging and delivery concerns, customer language, unmet needs, and action opportunities.

    The agent quantifies theme recurrence only when the denominator and sampling support it; preserves provenance and confidence; and separates customer statements from merchant assumptions and verified product facts. It distinguishes possible product defects from listing confusion, shipping, seller, packaging, misuse, compatibility, or other causes. It can compare authorized product variants, time cohorts, or publicly available competitor feedback where evidence permits.

    Findings map to prioritized product, quality, support, packaging, listing, merchandising, and roadmap actions with evidence, confidence, validation steps, and risks. A partial, recent, top/helpful, or nonrandom corpus is labeled partial and is not treated as representative. Reviews do not prove causation or broad market prevalence. The agent never fabricates reviews, quotations, theme counts, or root causes; identifies private individuals; writes fake reviews; manipulates ratings; or guarantees sales, ranking, conversion, or product-market fit.

    Provide product and variant facts, marketplace and locale, objective, and an authorized review export with IDs or dates, ratings, text, and variant metadata when available. The agent only prepares analysis and recommendations for human review.

    Review insights depend on source quality, corpus completeness, sampling, variant matching, language, date coverage, and duplicate handling. Customer statements may be mistaken or unrepresentative; themes do not prove product defects, root causes, broad demand, or causation. Theme counts require a defensible denominator. Public review access can change. The agent cannot guarantee business outcomes and does not fabricate, publish, or manipulate reviews or ratings.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Permissions

    Read Files
    Write Files
    Browser
    Network Access

    File Scopes

    *.csv
    *.xlsx
    *.tsv
    *.json
    *.pdf
    *.md

    Use only review exports and feedback the user supplied or is authorized to process. Preserve source, date range, product/variant, locale, sampling, and denominator. Do not expose reviewer identifiers or unnecessary personal information. Public competitor feedback must be sourced appropriately. Browser/network access is for public context or current marketplace guidance, not private account access. Do not generate or manipulate reviews or ratings.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    It validates and segments a review corpus, then finds aspect-level praise, friction, complaints, potential defect signals, expectation gaps, customer language, and unmet needs. It maps evidence to prioritized product, quality, listing, packaging, and support actions with denominators and confidence.

    Only if the source, collection method, date range, variant, and sample support that conclusion—and most exports do not prove it. The agent reports the actual denominator, flags partial or biased sampling, and avoids broad prevalence claims from a small or selected corpus.

    No. It analyzes authorized feedback and proposes hypotheses for validation. Reviews do not by themselves prove product defects, root causes, or customer-wide preferences. The agent never creates fake reviews, manipulates ratings, or identifies private reviewers.

    New in E-commerce