Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Search and read Wikipedia via x wkp — MediaWiki API, no API key, zero install; query, extract, suggest, and DDG route in one module. Load for wiki, wikipedia, encyclopedia lookup, article summary.
.claude/skills/x-cmd-wikipedia/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -48% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -65% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -61% | 0% |
Output is a plain-text extract (~first paragraph + infobox). Subcommand verb picks the shape — hop for search-then-summary, extract for known titles, search for candidates, suggest for did-you-mean. Default site is English Wikipedia; --lang / --api-url switch to other wikis.
x wkpbashx wkp hop Python # search + first-result extract in one shot x wkp extract OpenAI # summary of a known page x wkp search "Linux kernel" # list candidates when title is uncertain x wkp suggest pythen # did-you-mean: correction hints before search
x wkp -h for all flags and subcommands.
Upstream: https://en.wikipedia.org/w/api.php — MediaWiki action API. Sister projects (Wikidata, Wiktionary, Commons) are reachable the same way; point --api-url at the relevant subdomain.
For the questions that go one hop beyond a wikitext summary:
x wkp : <query> / x wkp ddgo <query> — same x wkp module, routed via DuckDuckGo's Wikipedia index (handles ambiguous or colloquial queries better than title search)x ddgo <query> — general web search, useful when Wikipedia has no article but the topic exists on the open webx hn / x rfc / x se research toolkitx wkp open <page> — fall back to the browser when you actually need the rendered page (infoboxes, tables, images)| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 16,489 | 3,022 | -82% | 1 | 1 | 0% | 1,165 | 656 | -44% | 0 | 0 | — |
case-02 | fail→pass | 9,009 | 4,941 | -45% | 1 | 1 | 0% | 1,249 | 655 | -48% | 0 | 0 | — |
case-03 | fail→pass | 22,868 | 4,819 | -79% | 1 | 1 | 0% | 1,649 | 669 | -59% | 0 | 0 | — |
case-04 | fail→pass | 21,300 | 2,194 | -90% | 1 | 1 | 0% | 1,646 | 571 | -65% | 0 | 0 | — |
case-05 | pass→pass | 14,240 | 3,853 | -73% | 1 | 1 | 0% | 1,018 | 710 | -30% | 0 | 0 | — |
case-06 | fail→pass | 11,420 | 3,891 | -66% | 1 | 1 | 0% | 1,745 | 684 | -61% | 0 | 0 | — |
case-07 | pass→pass | 18,182 | 78,855 | +334% | 1 | 1 | 0% | 1,473 | 798 | -46% | 0 | 0 | — |
case-08 | fail→pass | 7,680 | 2,328 | -70% | 1 | 1 | 0% | 874 | 633 | -28% | 0 | 0 | — |
case-09 | fail→pass | 27,207 | 3,043 | -89% | 1 | 1 | 0% | 1,438 | 692 | -52% | 0 | 0 | — |
case-10 | pass→pass | 19,620 | 3,216 | -84% | 1 | 1 | 0% | 1,078 | 617 | -43% | 0 | 0 | — |
case-11 | fail→pass | 10,462 | 4,272 | -59% | 1 | 1 | 0% | 1,586 | 816 | -49% | 0 | 0 | — |
case-12 | pass→pass | 206,171 | 3,650 | -98% | 1 | 1 | 0% | 1,810 | 697 | -61% | 0 | 0 | — |
case-13 | fail→pass | 13,417 | 3,446 | -74% | 1 | 1 | 0% | 1,979 | 801 | -60% | 0 | 0 | — |
case-14 | pass→pass | 18,440 | 2,777 | -85% | 1 | 1 | 0% | 970 | 663 | -32% | 0 | 0 | — |
case-15 | fail→pass | 12,695 | 2,551 | -80% | 1 | 1 | 0% | 1,434 | 688 | -52% | 0 | 0 | — |
case-16 | fail→pass | 13,339 | 2,365 | -82% | 1 | 1 | 0% | 2,124 | 656 | -69% | 0 | 0 | — |
case-17 | fail→pass | 10,845 | 7,727 | -29% | 1 | 1 | 0% | 1,330 | 677 | -49% | 0 | 0 | — |
case-18 | fail→pass | 13,967 | 3,641 | -74% | 1 | 1 | 0% | 2,326 | 931 | -60% | 0 | 0 | — |
case-19 | pass→pass | 7,017 | 3,868 | -45% | 1 | 1 | 0% | 1,025 | 671 | -35% | 0 | 0 | — |
case-20 | fail→pass | 7,114 | 2,638 | -63% | 1 | 1 | 0% | 851 | 693 | -19% | 0 | 0 | — |
case-21 | fail→pass | 32,601 | 3,192 | -90% | 1 | 1 | 0% | 1,382 | 737 | -47% | 0 | 0 | — |
case-22 | fail→pass | 11,145 | 4,011 | -64% | 1 | 1 | 0% | 1,613 | 903 | -44% | 0 | 0 | — |
case-23 | fail→pass | 8,902 | 3,442 | -61% | 1 | 1 | 0% | 1,003 | 624 | -38% | 0 | 0 | — |
case-24 | pass→pass | 6,290 | 3,097 | -51% | 1 | 1 | 0% | 1,046 | 710 | -32% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted. The headline lift of +71 percentage points is the difference between those two pass rates over the 24 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.