Verwendung
- Installationsbefehl kopieren und das Quellpaket vor der Installation in der eigenen Agentenumgebung prüfen.
- Benötigte Werkzeuge, Zugangsdaten und Hosts im Original prüfen.
- Eingabe und gewünschtes Ergebnis beschreiben und die Ausgabe mit den Quellen abgleichen.
npx skills add https://github.com/crawlora-org/crawlora-skills --skill journalist-media-researchVor dem Start
- Die Installation umfasst weder Modellabo noch API-Kontingent oder Dateizugriffsrechte.
- Sterne und Installationen zeigen Aktivität, keine unabhängige Qualitäts- oder Sicherheitsprüfung.
Aus dem Quell-Repository gerendert; Überschriften, Beispiele, Code, Tabellen, Links und Bilder bleiben erhalten.
Journalist and media research
Find reporters whose actual coverage fits a story, and explain the match. The output is a researched media list; sending pitches is a separate action requiring the user's request.
Setup and requests
Set CRAWLORA_API_KEY to your key from crawlora.net. Run the bundled scripts/crawlora.sh from this skill directory or by absolute path. It calls https://api.crawlora.net/api/v1 with x-api-key and prints JSON. Keep the key in the environment. Consult reference/endpoints.md for exact parameters and supported enum values.
Workflow
- Identify the story angle, relevant beats, target audience/geography, and
requested list size. A product category alone may need a more specific angle; use the user's supplied brief before asking for additional information.
- Inspect
/datasets/journalists/facetswithfacet=outlet,vertical, or
topic to discover the supported roster and filter values. Search by q, vertical, and/or returned exact topic/outlet values. q searches name, title, and bio, not article text. Use page_size up to 100, within the 10,000-record result window.
- Retrieve promising records from
/datasets/journalists/items/{outlet}/{slug}
using identifiers returned by search. Keep the outlet-specific identity: people can change outlets or contribute to more than one.
- Verify current affiliation and recent relevant work using
profile_url,
/bing/search, /bing/news, and /web/scrape. Search a name together with its outlet/domain to avoid namesakes. Confirm bylines and article dates on the source pages; search snippets alone are leads, not verified coverage.
- Rank by demonstrated story fit and recency. Cite one or two relevant pieces
per strong match when available. Mark candidates with only bio-level evidence as unverified rather than asserting they recently covered the topic.
scripts/crawlora.sh /datasets/journalists/facets facet=outlet vertical=tech
scripts/crawlora.sh /datasets/journalists/search vertical=tech page_size=10
# Refine with the story's terms; a topic label must come from facets:
scripts/crawlora.sh /datasets/journalists/search q=security page_size=10Output and coverage
Return a table or CSV with name, outlet, title, beat, profile_url, relevant_article_urls, article_dates, fit_reason, public_contact, contact_source_url, retrieved_at, and verification_status.
- The dataset covers a curated roster of supported outlets. An empty result
does not mean no journalist covers the topic. Use public search to supplement the roster and label externally discovered candidates.
- Beat topics are best-effort labels. Current bylined work is stronger evidence
than an old profile bio. Keep unknown affiliation or dates explicit.
contact_type=emailmeans an address was published on the outlet's page;
it does not mean the mailbox is verified or the person wants a pitch. social means only a social contact channel was found; none means neither. Do not invent addresses from an outlet's email pattern.
- Preserve contactless candidates when they are strong editorial matches,
unless contact availability is a requirement in the user's brief.
- Bound the search to the requested list size and explain any shortfall. On
429, back off; retry a transient 5xx once. Stop on 401/403, check the application code, and retain successful evidence if another source fails.
