An Agent Skill that teaches AI agents (Claude Code, Claude.ai, and other skill-compatible agents) how to use the free World Bank Indicators API.
The API exposes ~30,000 time-series indicators — GDP, population, inflation, unemployment, life expectancy, CO2 emissions, poverty, trade and more — for 217 economies and 78 aggregates, some series back to 1960. No API key, no signup.
Once loaded, an agent can reliably:
- Answer "GDP of X" / "compare Y across countries" in one step with the bundled
wb_data.pyscript (zero dependencies, Python stdlib only) - Query any indicator for any mix of countries, aggregates, or
alleconomies - Use the right time selectors (
dateranges,mrv,mrnev,gapfill) and know whymrnevbeatsmrvfor "the latest number" - Find indicator codes by keyword despite the API having no server-side search
- Discover indicators via topics (21) and sources (~70), and fetch country metadata (region, income level, capital)
- Pull the few monthly/quarterly series (Global Economic Monitor, quarterly debt)
- Get localized labels (
en,es,fr,ar,zh) - Avoid real-world pitfalls (XML default without
format=json, errors arriving as HTTP 200 message envelopes, archived indicator codes like the old CO2 series, aggregates mixed intocountry/all, economies the World Bank doesn't cover, transient HTML error pages)
All request/response examples and every indicator code in the skill were verified against the live API.
skills/
└── worldbank-api/
├── SKILL.md # Entry point: workflow, top indicator codes, errors, tips
├── scripts/
│ └── wb_data.py # Zero-dependency CLI: indicator + countries → formatted series
└── references/
├── endpoints.md # Full endpoint catalog + all query parameters
└── indicators.md # Verified indicator codes grouped by topic
This plugin is distributed through the NanookAI/skills marketplace. Inside Claude Code, run:
/plugin marketplace add NanookAI/skills
/plugin install worldbank-api@nanookai-skills
The skill is then available in every project, and /plugin marketplace update
picks up new versions.
Copy the skill folder into your project's .claude/skills/ directory:
cp -r skills/worldbank-api /path/to/your-project/.claude/skills/cp -r skills/worldbank-api ~/.claude/skills/Package the skill folder (respecting .skillignore) into a .skill bundle, or
upload the skills/worldbank-api/ folder contents wherever your agent platform
accepts skills. The skill is self-contained; the bundled script needs only a
Python 3.8+ standard library (no packages to install), and everything else is
plain documentation.
After installing, ask your agent things like:
- "Compare GDP per capita of Japan, Korea, and Germany over the last 20 years"
- "What's the current population of the world?"
- "Which countries have the highest inflation right now?"
- "Plot China's CO2 emissions per capita since 2000"
- "What share of people in India use the internet?"
- Base URL:
https://api.worldbank.org/v2(always addformat=json) - Free and keyless, no documented rate limit — but be gentle with concurrency; the server throttles bursts
- Official docs: Indicators API Documentation
- Data licensed under CC BY 4.0 (attribution: World Bank)
The skill documentation in this repository is provided as-is. World Bank data is subject to the World Bank's own terms of use.