ZooData — Amazon Seller Data Analysis

ZooData — Amazon Seller Data Analysis

Respond in the user's language. Use this skill for broad or composite Amazon research without a single specialized deliverable. Route a focused market-entry, market-discovery, or category-trend request to amazon-market-analysis; route a focused keyword, listing, pricing, review, or competitor request to its dedicated skill.

Start here

1. Classify the user's question. For a single data point, use the quick path in references/execution-guide.md; for multi-endpoint research, load its full workflow.
2. Read references/cli-contract.md before selecting or invoking a command. Inspect the bundled CLI's top-level and selected subcommand help.
3. Read references/reference.md for endpoint fields and local CLI options. Load the applicable scenario module: scenarios-composite.md for broad cross-validation, scenarios-eval.md for product evaluation, scenarios-expand.md for expansion, scenarios-listing.md for listing context, scenarios-ops.md for operations, or scenarios-pricing.md for pricing context.
4. Apply the shared interpretation and reporting rules in references/execution-guide.md; use the selected scenario for its specific evidence and conclusion scope.

Source-of-truth boundaries

• This SKILL.md owns trigger routing, module loading, the responsibility map, and non-negotiable runtime boundaries. It does not define product presets, thresholds, endpoint fields, procedures, or report templates.
• references/cli-contract.md owns shared CLI invocation, command identity, result classification, retries, and terminal or partial-result handling.
• references/reference.md owns endpoint availability, request and response fields, CLI parameter mapping, data shapes, and field identity. It does not set analytical thresholds.
• references/execution-guide.md owns shared workflow steps, product-preset selection, market-health interpretation, cross-endpoint analysis, and user-facing output rules. It does not redefine endpoint fields.
• Scenario modules own the evidence selection, conclusions, and scenario-specific report sections for their named task. They may narrow shared workflow rules but cannot redefine endpoint contracts or shared output rules.
• {skill_base_dir}/scripts/zoodata.py owns local preset expansion, request construction, transport metadata, and credential handling; it does not issue business conclusions.

Shared CLI Contract

Before selecting or invoking the first command, read and apply the local references/cli-contract.md. Reapply it after every granular or composite result and before any fallback, additional call, state write, interpretation, or report. The local references/execution-guide.md owns user-facing failure rendering.

Local Interface Failure Output

Use the failure rendering in references/execution-guide.md; direct a user with exhausted credits to https://zoodata.ai/en/pricing.

Non-negotiable boundaries

• Require ZOODATA_API_KEY; use the bundled CLI and the local command manifest. The credential-only check path precedes evidence calls. Stop and follow the shared CLI contract on credential, credit, validation, or terminal interface failures; do not invent missing evidence.
• Resolve and preserve a category path or ID before category-scoped comparisons. Distinguish inferred category paths from direct matches. Preserve marketplace, date, category scope, sample type, and the difference between full-category marketTotal and selected Top 100 marketSample metrics.
• Treat seller budget, experience, risk tolerance, and other profile text as local interpretation inputs. Send only the documented category/product identifiers and numeric filters to the API.
• State estimated credit cost before a broad multi-call scan. Require the user's explicit monitoring request before creating recurring work or saved baselines.

Capabilities and data flow

• Network: the bundled CLI sends authenticated requests to https://api.zoodata.ai and refuses an untrusted ZOODATA_BASE_URL.
• Execution: Python 3 standard-library shared CLI at {skill_base_dir}/scripts/zoodata.py.
• Local files: reads optional ~/.zoodata/config.json; review fallback may use a private temporary directory.
• Sent to the API: keywords, category paths/IDs, ASINs, marketplace, dates, and numeric filters. Seller profile text remains local.
• Credits: evidence calls consume account credits; the credential-only check path does not.

Execution entry

python {skill_base_dir}/scripts/zoodata.py <documented-subcommand> ...

This skill allows categories, market, market-structure-profile, market-history, products, competitors, product, analyze, report, opportunity, history, brand-overview, brand-detail, price-band-overview, price-band-detail, check, reviews-raw, review-tag-prompt, review-reduce-prompt, and review-aggregate.