wikipedia

Use this skill when the user asks for Wikipedia facts, summaries, related pages, or media links.

Workflow:

1. •Pick language (default en).
2. •Normalize titles:
   • •trim spaces
   • •replace spaces with _
   • •URL encode when needed
3. •If the title is uncertain → search first.
4. •Use action_run_command with curl -s <url>.
5. •Retry once using search if a page endpoint fails.
6. •Return concise results with the final Wikipedia URL.

Supported endpoint patterns:

• •Search titles: https://{language}.wikipedia.org/w/rest.php/v1/search/title?q={query}&limit={limit}
• •Summary: https://{language}.wikipedia.org/api/rest_v1/page/summary/{title}
• •Full text extract (JSON): https://{language}.wikipedia.org/w/api.php?action=query&format=json&prop=extracts&explaintext=1&redirects=1&titles={title}
• •Related pages: https://{language}.wikipedia.org/api/rest_v1/page/related/{title}
• •Media list: https://{language}.wikipedia.org/api/rest_v1/page/media-list/{title}
• •Random summary: https://{language}.wikipedia.org/api/rest_v1/page/random/summary

When to use each endpoint:

• •Search titles → when user gives a topic/name but title may be wrong or ambiguous.
• •Summary → default for “who is / what is”.
• •Full text extract → when user asks for full article or detailed info.
• •Related pages → when user asks for similar or connected topics.
• •Media list → when user asks for images/media.
• •Random summary → when user asks for a random Wikipedia article.

Input expectations:

• •Search requires query; optional limit (default 5).
• •Page endpoints require title.

Canonical article link format: https://{language}.wikipedia.org/wiki/{Title_With_Underscores}

Error handling:

• •If HTTP status is not 200:
  1. •Run title search using the same input.
  2. •Retry the page endpoint using the top result title.
  3. •If still failing, return a short failure reason and show search results.