Free, keyless, liveness-verified APIs for real data - Claude executes instead of deliberating.
Reach for this when a task needs real data - weather, geocoding, FX and crypto, earthquakes and NASA imagery, country statistics, dictionary and Wikipedia lookups, trivia, or realistic fake data for a demo - and you want Claude to hit a working free API on the first try instead of deliberating over a directory or reaching for key-walled defaults like OpenWeatherMap and Google Maps from training memory. Each skill names one ranked default per need with exact URL templates, real response shapes with field names and units, and the rate-limit and encoding gotchas that break naive integrations. Every endpoint is probed live (HTTP 200, parseable JSON, under 2s) by a rerunnable health check committed alongside the skills; dead and key-walled APIs are called out by name so they never get cited. A router skill owns the vague 'I need real data for X' trigger and dispatches to the right domain skill. Sourced and liveness-verified from the public-apis project (MIT).
Arranged in the author's recommended order. Walk through them in sequence, or open any one on its own.
Use when a task needs real, live data but the domain is vague or unstated - "I need real data for X", "is there a free API for this?", "pull actual data into this demo", "no mock data, use a real source", or picking a data source before building. Routes to the right Live Data skill (weather-climate, geo-places, finance-fx, space-earth-science, government-open-data, test-placeholder-data, language-reference, fun-content), every one backed by keyless liveness-verified APIs. Do NOT use when the domain is already explicit - a request naming weather, FX, earthquakes, or fake users goes straight to that skill; do NOT use for designing your own API - use api-design instead.
View skill