Scoped source access review — 5 October 2026

Open AI agent: Researcher. This is the project's documented acquisition assessment; it is not a publisher approval, a legal opinion or a blanket reuse licence. Activation is established by the runtime manifest and a verified data release.

NASA: one factual programme overview

• Exact source: NASA Internship Programs. Direct public GET returned HTTP 200 without redirect on 5 October 2026. The canonical WebPage JSON-LD identifies the same URL and original title. Original datePublished is 2023-01-23T11:57:52-05:00; dateModified is 2026-09-28T15:01:36-04:00. Modification must not replace publication.

• NASA robots returned HTTP 200 and permits all paths for the wildcard agent; no wildcard crawl delay was stated. Proposed collection is one public request per six-hour run, without authenticated access or bypasses.

• NASA content-use guidelines permit factual informational use without implying endorsement, with source acknowledgment. This project review covers deterministic copying of the original title, canonical link and original publication timestamp only. It excludes prose, imagery, marks, identifiable persons and third-party material.

• The policy's AI section distinguishes factual disclosure of source material from generated outputs and prohibits implying NASA approval of an AI product. This adapter does not generate a summary, train a model, or attribute AI-generated statements to NASA. Source disclosure must not imply publisher review, permission for a specific AI use or endorsement.

• Keep summary empty; use internships and programme-overview; leave availability, deadline, host country and eligible countries unknown. The overview combines OSTEM and Pathways, so no individual vacancy or single deadline is established by this record. Each application requires checking the original programme. No images or NASA branding are imported.

• The attempted programme /feed/ URL returned HTML rather than RSS. The general agency feed is mixed news and provides no selected opportunity here. A trusted exact-page HTML adapter is the appropriate scoped acquisition method.

IKY: explicit permission required

• The actual footer-linked IKY terms were located and reviewed on 5 October 2026. The prior guessed terms URL returning 404 did not establish publisher policy.

• Section 3 requires prior permission for public copying, reproduction, republication, storage and translations, including partial or summarized content. Its exception is strictly personal computer use, which does not authorize this public catalogue. Section 8 permits ordinary hyperlinks without permission and disallows framing.

• IKY robots returned HTTP 200 and allows the public feed; RSS still parses 24 items. Technical access does not override the explicit republication restriction. Keep the adapter disabled with permission_required; no publisher permission has been obtained.

• Most feed items are results, administrative notices or institutional announcements. Selection would remain necessary even after permission is obtained. The record's publisher activity evidence does not claim that every feed item is an open youth application.

Portugal Fulbright: one exact master's programme overview

• The master's programme, robots, homepage and published privacy page each returned HTTP 200 on 5 October 2026. Canonical and Open Graph URL identify the same programme, with original Open Graph title Bolsa Fulbright para Mestrado.

• Robots allows the public programme path and prohibits /wp-admin/; no crawl delay is stated. Collection makes one programme-page request per six-hour pipeline run, using the project User-Agent and existing timeout/size/no-redirect bounds. It accesses no application portal or personal data.

• The privacy page concerns candidate/recipient personal information and confidentiality. The site retains an all-rights-reserved footer. No express factual-title/link indexing prohibition or reuse licence was established in the reviewed public pages; no publisher approval or endorsement is claimed. The project's scoped acquisition assessment permits deterministic indexing of the factual programme title and original link only, excluding descriptions, imagery, logos, documents and personal data. This assessment does not grant a licence to reproduce source content; new restrictive terms require suspension/review.

• The HTML provides article:modified_time 2026-10-01T14:33:53+00:00, but no original publication date. Keep published_at null; do not substitute modification, first-seen time or an application opening date for publication. The trusted Open Graph selection requires exactly one canonical, matching og:url and nonempty og:title; missing/ambiguous identity fails and preserves prior state.

• The reviewed programme describes 2027/28 master's funding and a 1 October 2026–15 January 2027 application window, closing 23:59 Lisbon time; Portuguese citizenship and additional exclusions apply. These facts are editorial research evidence, not encoded as a generic recurring programme's application availability/eligibility. The catalog record remains scholarships, programme-overview, empty summary and unknown status/deadline/destination/eligible countries; PT is publisher geography. Always consult the original publisher for current application conditions.

• The general Portugal RSS remains excluded because its inspected news items do not contain this programme; enabling the exact programme-page manifest does not authorize broader feed ingestion. This adds one reviewed adapter, not three operational sources per country. Production collection/release and site acceptance must be checked separately.

Netherlands Fulbright: one exact doctoral programme overview

• The doctoral programme, robots, privacy declaration and homepage returned HTTP 200 without redirect on 5 October 2026. The direct review used the clear project User-Agent and at least ten seconds between same-host requests. Robots states User-agent: * and Crawl-delay: 10, with no disallow rule; it is access evidence, not a reuse licence. One public programme request per six-hour collection stays within that observed rate.

• The linked privacy declaration concerns applicant, recipient and visitor personal data. Its consent requirements apply to personal records and reference letters, which this adapter never accesses. No express factual-title/link indexing prohibition or reuse licence was established in the exact programme, homepage and privacy scope. The project's existing scoped metadata assessment supports only deterministic original title and canonical link, with no publisher approval, endorsement or blanket copyright permission claimed. Newly restrictive terms or access blocks require suspension/review.

• Direct current HTML contains exactly one canonical link and matching og:url; the sole og:title is fulbright beurzen voor promovendi. No original publication metadata exists. Reuse the tested Open Graph selection; published_at stays null and collection time is not publication. Retain only sanitized identity tags in fixtures.

• The programme serves doctoral researchers affiliated with Dutch institutions; nationality restrictions apply. Its current 2027/28 round closes 1 December 2026 at noon Dutch time. These are date-stamped research facts, not encoded into the recurring catalogue record: keep scholarships/programme-overview, empty summary, unknown availability/deadline and empty host/eligible countries. NL is publisher geography only.

• General Netherlands RSS stays excluded for irrelevant alumni/blog items. This exact page is a separate bounded integration; local validation and enabled configuration alone do not establish production publication or website acceptance. No application form, programme PDF, prose, image or personal information is imported.

Luxembourg FNR AFdoc — exact programme metadata, 5 October 2026

Direct requests to the programme, robots and actual footer-linked legal notice returned HTTP 200 without redirect. Robots has an empty global Disallow and lists a sitemap. One programme request per six-hour collection is conservative; no other FNR page is collected at runtime.

The legal notice expressly permits information reproduction for informational purposes, with express consent required for commercial purposes, and hyperlinks that clearly preserve the source origin. Scientific-reference reuse additionally requires the notice © Fonds National de la Recherche. This adapter publishes no scientific-reference text, article prose, images, logos or third-party content; its bounded informational index preserves the original programme title, canonical link and original publication date. The publisher name/link identifies the origin. No FNR endorsement, general reuse licence or publisher approval is claimed. Source terms may change and remain authoritative.

The programme's exact canonical URL has one matching JSON-LD WebPage: title AFdoc - FNR, original publication 2026-07-29T13:25:54+00:00, modification 2026-09-16T10:28:12+00:00. Additional conflicting Open Graph titles are present; the existing exact JSON-LD adapter is reused, with no relaxed identity matching. Modification is never substituted for publication. The sanitized fixture contains identity/date metadata only.

The catalogue record is scholarships/programme-overview. Summary stays empty, application status unknown, deadline null and host/eligible countries empty. Luxembourg describes publisher geography only. Editorial evidence of a November 25 doctoral funding call does not change these machine-extracted fields or establish unrestricted youth eligibility. Local tests/direct collection support implementation readiness; connected/production claims require a successful merged-main producer release and independently verified asset.