Skip to content

[ARTICLE] 2026-04-06 #21

@meinstein

Description

@meinstein

Required Inputs

Public URL

https://www.nytimes.com/2026/04/05/us/politics/stephen-miller-immigration-agenda.html

Archive URL

https://archive.ph/Boohc

Optional Inputs

Private URL

ADD_PRIVATE_URL_HERE

Notes

ADD_NOTES_ABOUT_ARTICLE_HERE


Copilot Task

Use this issue to create exactly one new context.yml file.

Priority of truth:

  1. Explicit field values in this issue
  2. Optional notes
  3. Values inferred from the fetched archive content

If sources conflict, prefer the higher-priority source.

LLM Content Fetching and Field Inference

Fetch the article content from the Archive URL. Use the fetched content to infer any fields not explicitly provided in this issue, including:

  • title: infer from the article's <title> tag or main heading
  • language: infer from the content's language (use BCP 47 codes, e.g. en, fr)
  • date: infer from the article's published/updated date metadata or byline; set to null if not present
  • website: infer from the archive URL's origin domain (e.g. example.com)
  • tags: analyze the main body of the article and generate a list of relevant tags that describe what the article is about

Tags should be concise, lowercase, and reflect the main topics, entities, or themes present in the article.

Output Contract

Generate one YAML file using this template:

type: article
language: <from Language field, or inferred from article content>
author: <from Author field, or null>
date: <from Date field, or inferred from article content, or null>
title: <from Title field, or inferred from article content>
website: <from Website field, or inferred from archive URL>
url: <from Public URL field, or null>
archive_url: <from Archive URL field, or null>
private_url: <from Private URL field, or null>
notes: <from Notes field, or null>
tags:
	- tag1
	- tag2
	- ...

Explicit field values in this issue always take precedence over inferred values. If a field cannot be provided or inferred, set it to null. If no tags are generated, set tags: [].

Save file as: <YYYY>/<MM>/<DD>/context/<unix_timestamp_of_issue_created_at>/context.yml

Where:

  • YYYY/MM/DD is based on the date in the issue title
  • unix_timestamp_of_issue_created_at is based on the GitHub issue creation timestamp

Then:

  1. Create missing directories (if a context.yml already exists at that location, stop and comment).
  2. Commit and open a PR.
  3. Use the repository default branch as PR base (do not assume main).
  4. Do not ask follow-up questions unless blocked.

Metadata

Metadata

Labels

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions