amazon-validationlisted

# Amazon Validation This Skill does not fetch any new data itself. It teaches you how to correctly read what `hunt discover --with-market-data` and `hunt show <opportunity_id>` already produced, so evidence isn't over-trusted, under-trusted, or silently reclassified in your report. Full rule detail lives in `references/evidence-model.md` (what tier is what) and `references/validation.md` (what can go wrong with a live provider call and how the engine already handles it). This file is the operational checklist for using those two together. ## 0. Before touching any real-mode command Run `hunt providers`. It shows, per provider, exactly which of the two gating conditions (credential present, `terms_acknowledged`/ `license_acknowledged` set) is missing — see `validation.md` section 2 (D3, D4). Never guess whether a provider is usable; always check this first. ## 1. Reading `hunt discover --with-market-data <mode>` output The printed summary includes live-enrichment call/failure/skip counts. Read them literally: - **call** = a live provider attempt was made and returned data. - **failure** = `ProviderRequestError` fired (`validation.md` D2) — that specific opportunity was **not evaluated**, not "found to be weak." Report it as "validation attempt failed," never as a rejected candidate on the merits. - **skip** = the candidate never reached live enrichment (cheap filter already eliminated it before spending a provider call) — not "provider tried and found nothing.