When to use this skill
When the user asks "what is X", "tell me about X", or wants a quick Wikipedia overview — summaries, article descriptions, thumbnails, or topic extracts. This is the fastest path from a topic name to a structured answer: title, Wikidata description, plain-text extract, and thumbnail in one unauthenticated GET. For full article HTML (100 KB+ payloads), go directly to the /page/html/{title} endpoint — this skill stops at the summary.
Your best first call
curl "https://en.wikipedia.org/api/rest_v1/page/summary/Fermi_paradox"
No auth. No key. Returns a JSON object. Use /page/summary/{title} when you know the exact article title — underscores for spaces, parenthetical disambiguation where needed (e.g. Python_(programming_language)). When you only have a keyword or vague topic, search first:
curl "https://en.wikipedia.org/w/api.php?action=opensearch&search=fermi+paradox&limit=5&format=json"
OpenSearch returns a 4-element array: [query, [titles], [descriptions], [urls]]. The descriptions array is often empty strings — do not rely on it for disambiguation. Pick the right title from the titles array, then pass it to the summary endpoint.
Key fields from the summary response:
type — "standard" for articles, "disambiguation" for disambiguation pages. Check this before using the extract.title — article titledescription — Wikidata short description, editor-curated metadata that disambiguates the topic at a glanceextract — plain-text article extract, the payload you relay to the userthumbnail.source — hosted image URLcontent_urls.desktop.page — link to the full Wikipedia articlewikibase_item — Wikidata entity ID for structured cross-references beyond Wikipedia
Fallbacks (when the best call isn't enough)
- User wants a random article →
https://en.wikipedia.org/api/rest_v1/page/random/title returns a random article's title and edit metadata. Pass the returned title to the summary endpoint for the actual content.
- User needs the full article body →
/api/rest_v1/page/html/{title} returns complete HTML. Expect 100 KB to over 1 MB — only use when the summary is genuinely insufficient for the user's question.
Pitfalls
- Disambiguation pages return
"type": "disambiguation" with an extract that is a list of meanings, not an answer. Always check type before using extract. If you hit one, search OpenSearch with a more specific query.
- Article titles use underscores and parenthetical disambiguation: "Mercury" is a disambiguation page covering the planet, element, car brand, and musician; "Mercury_(planet)" is the article. Users will rarely give you the exact title — search first when in doubt.
- The Action API (
/w/api.php) returns rendered HTML, not JSON, if you forget format=json. Always include it.
- The
/page/html/{title} endpoint returns 100 KB+ responses. Do not call it reflexively when the summary would do.
One-line summary for the user
I can fetch Wikipedia article summaries and search for article titles — no auth, one GET per call — but I need the exact title with underscores and disambiguation suffixes, or I'll hit a disambiguation page.
SKILL.md source (frontmatter + body)
---
name: access-wikipedia-article-summaries
description: When the user asks "what is X", "tell me about X", or wants a Wikipedia summary, article overview, topic extract, or thumbnail — reach for the Wikipedia REST API. No auth, one GET per call.
---
## When to use this skill
When the user asks "what is X", "tell me about X", or wants a quick Wikipedia overview — summaries, article descriptions, thumbnails, or topic extracts. This is the fastest path from a topic name to a structured answer: title, Wikidata description, plain-text extract, and thumbnail in one unauthenticated GET. For full article HTML (100 KB+ payloads), go directly to the `/page/html/{title}` endpoint — this skill stops at the summary.
## Your best first call
```bash
curl "https://en.wikipedia.org/api/rest_v1/page/summary/Fermi_paradox"
```
No auth. No key. Returns a JSON object. Use `/page/summary/{title}` when you know the exact article title — underscores for spaces, parenthetical disambiguation where needed (e.g. `Python_(programming_language)`). When you only have a keyword or vague topic, search first:
```bash
curl "https://en.wikipedia.org/w/api.php?action=opensearch&search=fermi+paradox&limit=5&format=json"
```
OpenSearch returns a 4-element array: `[query, [titles], [descriptions], [urls]]`. The `descriptions` array is often empty strings — do not rely on it for disambiguation. Pick the right title from the `titles` array, then pass it to the summary endpoint.
Key fields from the summary response:
- `type` — `"standard"` for articles, `"disambiguation"` for disambiguation pages. Check this before using the extract.
- `title` — article title
- `description` — Wikidata short description, editor-curated metadata that disambiguates the topic at a glance
- `extract` — plain-text article extract, the payload you relay to the user
- `thumbnail.source` — hosted image URL
- `content_urls.desktop.page` — link to the full Wikipedia article
- `wikibase_item` — Wikidata entity ID for structured cross-references beyond Wikipedia
## Fallbacks (when the best call isn't enough)
- **User wants a random article** → `https://en.wikipedia.org/api/rest_v1/page/random/title` returns a random article's title and edit metadata. Pass the returned `title` to the summary endpoint for the actual content.
- **User needs the full article body** → `/api/rest_v1/page/html/{title}` returns complete HTML. Expect 100 KB to over 1 MB — only use when the summary is genuinely insufficient for the user's question.
## Pitfalls
- Disambiguation pages return `"type": "disambiguation"` with an extract that is a list of meanings, not an answer. Always check `type` before using `extract`. If you hit one, search OpenSearch with a more specific query.
- Article titles use underscores and parenthetical disambiguation: "Mercury" is a disambiguation page covering the planet, element, car brand, and musician; "Mercury_(planet)" is the article. Users will rarely give you the exact title — search first when in doubt.
- The Action API (`/w/api.php`) returns rendered HTML, not JSON, if you forget `format=json`. Always include it.
- The `/page/html/{title}` endpoint returns 100 KB+ responses. Do not call it reflexively when the summary would do.
## One-line summary for the user
I can fetch Wikipedia article summaries and search for article titles — no auth, one GET per call — but I need the exact title with underscores and disambiguation suffixes, or I'll hit a disambiguation page.