---
name: legal-document-source-research
description: "Research and verify authoritative, legally downloadable editions, scans, translations, and manuscript facsimiles."
version: 1.0.0
author: Hermes Agent
license: MIT
metadata:
  hermes:
    tags: [research, legal-sources, public-domain, open-access, archives, manuscripts]
---

# Legal Document Source Research

Use when a user asks for authoritative and legal PDFs/downloads of historical texts, translations, critical editions, archival scans, manuscripts, or other primary-source documents.

## Source hierarchy

1. Public-domain scans contributed by named libraries to Internet Archive or comparable institutional archives.
2. University/institutional repositories with stable DOI, URN, or handle records.
3. Official holding institutions, antiquities authorities, museums, or manuscript libraries for facsimiles and catalog metadata.
4. DOAB/OAPEN and publisher-hosted open-access editions with explicit licenses.
5. Commercial reposts, personal sites, and unsourced PDFs are discovery leads only, not final evidence.

## Workflow

1. Identify the exact work and distinguish recensions, translations, editions, reprints, critical texts, studies, and facsimiles.
2. Search authoritative repositories first; use repository APIs and metadata endpoints when ordinary search is noisy or blocked.
3. Verify the exact file URL and whether it is truly downloadable, unencrypted, and public—not controlled lending.
4. Extract edition metadata from the catalog and, where needed, the scanned title page/OCR.
5. Inspect the PDF contents when necessary to establish language, pagination, coverage, and whether it contains a full translation or only excerpts.
6. Verify the legal basis: explicit OA license, repository authorization, public-domain basis, or restrictive terms of use.
7. Return a stable landing page and exact direct download URL, metadata, source authority, legal status, and limitations.

## Required distinctions

Never collapse these categories:

- manuscript images/facsimile;
- transcription;
- English translation;
- critical edition with translation;
- English-language study that quotes translated passages;
- complete standalone translation.

A facsimile is not a translation. An English monograph discussing a text is not automatically a complete translated edition.

## Internet Archive verification

Prefer repository APIs over search-result snippets:

- Advanced Search: `https://archive.org/advancedsearch.php?q=...&fl[]=identifier,title,creator,date,format&rows=50&output=json`
- Metadata: `https://archive.org/metadata/IDENTIFIER`
- Item: `https://archive.org/details/IDENTIFIER`
- Download: `https://archive.org/download/IDENTIFIER/FILENAME.pdf`

Inspect the metadata `files` array to confirm the exact filename, format, and access state. Prefer unencrypted `Text PDF` or image-container scans from named institutional contributors. `ACS Encrypted PDF` and `LCP Encrypted PDF` indicate lending/access controls and must not be presented as open PDF downloads.

When catalog metadata is incomplete, inspect `IDENTIFIER_djvu.txt` or the scanned title page. Confirm whether edition language such as “second edition” reflects a genuine revision or merely a reprint.

## Legal-status discipline

- Publicly reachable does not automatically mean public domain or openly licensed.
- For older editions, state the basis—publication year and relevant author death year—and qualify public-domain conclusions by jurisdiction.
- If an institutional repository permits unrestricted download but shows no Creative Commons license, describe it as repository-authorized access; do not assume permission to redistribute or adapt.
- For official image repositories, inspect and cite reuse terms. Private-use access may be legal while republication still requires permission.
- If no legal open full translation can be verified, say so. Provide the best official facsimile and OA scholarship, explicitly labeling what each lacks.

## Discovery fallbacks

- Crossref: exact title, author, publisher, publication date, DOI, ISBN.
- OpenAlex: repository locations, `best_oa_location`, PDF URL, and license indicators.
- DOAB/OAPEN: OA book metadata and license verification.
- DOI redirect or repository DOM: direct PDF link discovery.
- Local `pdfinfo` and `pdftotext`: verify page count, language, contents, and translation coverage.

See `references/ancient-jewish-text-editions.md` for a verified example package covering 1 Enoch, Jubilees, Book of Giants, and the War Scroll.

See `references/modern-copyrighted-scripture-downloads.md` for modern copyrighted Bible translations: distinguish browser access, offline app downloads, and exportable files; verify canon scope and resolve official store links.