pangolinfo-amazon-product-explorer

The name/description promise Amazon product discovery via the Pangolinfo data engine and the bundled scripts make outbound calls to pangolinfo endpoints (scrapeapi.pangolinfo.com) matching that purpose. Minor inconsistency: registry metadata lists PANGOLINFO_API_KEY, PANGOLINFO_EMAIL, and PANGOLINFO_PASSWORD as 'required', while SKILL.md and the scripts indicate an API key OR email+password are acceptable (API key is recommended). This is a documentation/metadata mismatch, not a capability mismatch.

SKILL.md SOP and the scripts limit actions to querying Pangolinfo APIs, parsing results, and optional local caching. The SOP mandates language detection and defaults (Amazon US, zip 90001). There are no instructions to read unrelated files, harvest system secrets, or transmit data to endpoints other than Pangolinfo. The scripts reference a per-user cache file (~/.pangolinfo_api_key) but caching is opt-in and currently disabled by default in the code.

No install spec or external downloads; the package is a flat set of zero-dependency Python scripts using the standard library (urllib). Nothing is fetched from arbitrary URLs or written to nonstandard system locations by default.

The only credentials requested are Pangolinfo credentials (API key or email+password), which are proportional to the described functionality. The minor issue: registry metadata implies all three env vars are required, whereas SKILL.md and the clients accept either an API key OR email+password. Users should avoid pasting credentials into public places and prefer using an API key.

The skill is not forced-always and uses normal autonomous-invocation defaults. It does not request elevated system privileges. Scripts may optionally persist an API key to ~/.pangolinfo_api_key if caching is explicitly enabled; by default the code disables disk caching (CACHE_TO_DISK = False).