Sources: Wikimedia Reference Reliability Lists

This page lists every source list aggregated into WikiSignals' reference-reliability dataset, what each one contains, and where the raw data comes from. The dataset is rebuilt monthly; individual list contents change as Wikipedia editors update the underlying pages.

87 distinct source lists are fetched across the 10 source modules below (count computed from each module's actual page-fetch list, not a manual estimate). This excludes the Chinese Perennial Sources list's 366 internal per-entry sub-pages, which are structural components of that ONE list rather than separate lists -- see the note under Reliable/Perennial Sources below.

Two things get called "a source" below, and it's worth keeping them apart. Most rows are direct sources: the wiki page or file we fetch is the curated list -- an editor (or a template) maintains it, and we just read it. Cite Unseen is different: it is a user script (a browser tool) whose own code contains no domain data at all -- at runtime it fetches and parses ~47 separate wiki pages (Cite_Unseen/sources/*), each independently maintained and each a direct source in its own right. We fetch those same 47 pages ourselves and never run Cite Unseen's script; it's grouped here as one cite_unseen row purely for bookkeeping convenience, not because it's a single list. (UPSD and CiteHighlighter are also hosted as .js files, but unlike Cite Unseen, their domain data is embedded directly in that one file -- no further pages to fetch -- so we still count them as direct sources.)

Source # lists What it contains Representative URL(s)
Wikipedia fake-news / disinformation / satire lists (en, cs, ar) (fakenews) 12 domain + category (one category per source page, e.g. "fake news website", "troll farm") + name (the table's Name column, when present) https://en.wikipedia.org/wiki/List_of_fake_news_websites
https://en.wikipedia.org/wiki/List_of_corporate_disinformation_website_campaigns
https://en.wikipedia.org/wiki/List_of_political_disinformation_website_campaigns
(+9 more shown in appendix)
Wikipedia Citation Watchlist (citation_watchlist) 4 domain + category (page name) + warning_label (level-2 heading within the page, e.g. Warn/Caution/Inspect) https://en.wikipedia.org/wiki/Wikipedia:Citation_Watchlist/Lists/Self-Publishing
https://en.wikipedia.org/wiki/Wikipedia:Citation_Watchlist/Lists/Predatory_Open_Access_Journals
https://en.wikipedia.org/wiki/Wikipedia:Citation_Watchlist/Lists/Wikimedia
(+1 more shown in appendix)
User:Kuru/fakesources (curated list of paid-placement / fake-source websites) (kuru) 1 domain + category (section heading, e.g. "Main list", "Blacklisted sites") + name (an adjacent [[wikilink]] to the outlet, present on only a small minority of entries) https://en.wikipedia.org/wiki/User:Kuru/fakesources
Wikimedia spam blocklists (BlockedExternalDomains + regex Spam-blacklist) (spam_blacklist) 4 domain only (no category) -- BlockedExternalDomains entries carry the literal domain; Spam-blacklist regex lines are expanded into literal domains per the bounded-alternation rule (unbounded/char-class lines are skipped, logged separately) https://en.wikipedia.org/wiki/MediaWiki:BlockedExternalDomains.json
https://meta.wikimedia.org/wiki/MediaWiki:BlockedExternalDomains.json
https://en.wikipedia.org/wiki/MediaWiki:Spam-blacklist
(+1 more shown in appendix)
Reliable/Perennial Sources (RSP), 15 language editions (perennial_sources) 15 domain + category (per-language reliability status, e.g. "Generally reliable" / "Generally unreliable" / "Deprecated" / "Blacklisted"; bg/fr have no status -- list membership alone is the signal). Note: the ru, el, simple-en, and tr wiki equivalent pages list source name plus status only, with no domain column at all -- verified by direct inspection, not a parsing bug -- so those 4 languages contribute 0 domain rows despite being fetched (kept in the codepage list for completeness / in case the page structure changes). https://en.wikipedia.org/wiki/Wikipedia:Reliable_sources/Perennial_sources
Unreliable/Predatory Source Detector (User:Headbomb/unreliable.js) (upsd) 1 domain (or doi_prefix) + category (the rule's comment field, reused across several rules) https://en.wikipedia.org/wiki/User:Headbomb/unreliable.js
CiteHighlighter (User:Novem Linguae/Scripts/CiteHighlighter) (citehighlighter) 1 domain + category (the JSON object's rating key, e.g. "generallyUnreliable") https://en.wikipedia.org/wiki/User:Novem_Linguae/Scripts/CiteHighlighter/SourcesJSON.js
Cite Unseen source lists (cite_unseen) 47 domain + category (normalized parent category, e.g. "advocacy", "news", "predatory"; WikiProject/RSP-style pages instead carry a {source}{Tier} label, e.g. "enRspBlacklisted") + name (the CULink template's pub= param, present on only a minority of entries) https://meta.wikimedia.org/wiki/Cite_Unseen/sources
New Pages Patrol source guide (npp_guide) 1 domain + category ("{region}/{country}" or "{topic}") + status (Reliable / Unreliable / No consensus) + name (outlet name, from a preceding [[wikilink]] or plain text before the domain link) https://en.wikipedia.org/wiki/Wikipedia:New_pages_patrol_source_guide
Deprecated sources (domains list) (deprecated_domains) 1 domain + category (fixed: "Deprecated") + name (outlet name, rowspan-carried across multiple domain rows) https://en.wikipedia.org/wiki/Wikipedia:Deprecated_sources/Domains
Total 87

Output format

The aggregator publishes one JSON file (wiki_ref_reli_domains.json) keyed by normalized domain. Each domain entry lists every source that flagged it, plus that source's category label for it where the source has categories:

{
      "breitbart.com": {
          "sources": [
              "upsd",
              "deprecated_domains",
              "perennial_sources"
          ],
          "names": [
              "Breitbart News"
          ],
          "hits": [
              {
                  "source": "upsd",
                  "category": "Blacklisted source"
              },
              {
                  "source": "deprecated_domains",
                  "category": "Deprecated",
                  "name": "Breitbart News"
              },
              {
                  "source": "perennial_sources",
                  "category": "Blacklisted",
                  "name": "Breitbart News",
                  "lang": "en"
              }
          ]
      }
  }
  

A domain appearing under many sources is not necessarily "worse" than one appearing under one -- some sources (e.g. Cite Unseen's RSP mirror, Citation Watchlist's RSP list, and the Perennial Sources page itself) independently track overlapping ground, so multi-source hits partly reflect redundant coverage rather than independent corroboration. Check the hits array and the categories involved before treating source count alone as a reliability signal.

Not every source supplies a human-readable site name -- Citation Watchlist, the spam blocklists, UPSD, and CiteHighlighter are domain-only, with no name field anywhere in their source pages. domains[domain].names collects whatever name(s) the other sources did supply; it can be empty even for a well-attested domain if only name-less sources flagged it.

Appendix: every list URL fetched

Full enumeration behind the counts above -- one list per line, grouped by source module.

Wikipedia fake-news / disinformation / satire lists (en, cs, ar) (fakenews) -- 12 lists

Wikipedia Citation Watchlist (citation_watchlist) -- 4 lists

User:Kuru/fakesources (curated list of paid-placement / fake-source websites) (kuru) -- 1 list

Wikimedia spam blocklists (BlockedExternalDomains + regex Spam-blacklist) (spam_blacklist) -- 4 lists

Reliable/Perennial Sources (RSP), 15 language editions (perennial_sources) -- 15 lists

Unreliable/Predatory Source Detector (User:Headbomb/unreliable.js) (upsd) -- 1 list

CiteHighlighter (User:Novem Linguae/Scripts/CiteHighlighter) (citehighlighter) -- 1 list

Cite Unseen source lists (cite_unseen) -- 47 lists

New Pages Patrol source guide (npp_guide) -- 1 list

Deprecated sources (domains list) (deprecated_domains) -- 1 list

(The Chinese Perennial Sources list above is 1 of the 15 perennial_sources entries; its 366 individual per-source infobox sub-pages, one per outlet, are fetched in bulk as part of reading that single list and are not separately enumerated here.)