# AI security digest JSON

`https://jonathanglass.cloud/ai-security.html` is the HTML page.

`https://jonathanglass.cloud/data/ai-security/digest.json` is the dataset.

```json
{
  "schemaVersion": 1,
  "windowDays": 400,
  "description": "How to read the file",
  "sources": [{"id": "aws-whats-new", "name": "AWS What's New", "feedUrl": "https://...", "siteUrl": "https://...", "note": "..."}],
  "sourceStatus": {"aws-whats-new": {"ok": true}},
  "items": [{
    "title": "Publisher title",
    "url": "https://publisher.example/item",
    "sourceId": "aws-whats-new",
    "sourceName": "AWS What's New",
    "published": "YYYY-MM-DD",
    "summary": "Publisher excerpt",
    "summaryTruncated": false
  }]
}
```

`title`, `published`, `summary`, and `url` are copied from the feed. `summary` is plain text.
`summaryTruncated` is true when the excerpt was cut to 480 characters. The file does not add a paraphrase.

A day page is `https://jonathanglass.cloud/ai-security/YYYY-MM-DD.html` (Markdown at the `.md` twin).
RSS: `https://jonathanglass.cloud/ai-security/feed.xml`. Atom: `https://jonathanglass.cloud/ai-security/atom.xml`.

`scripts/fetch_ai_security.py` reads the feeds and adds new URLs. It does not need an API key.
`scripts/render_machine_readable.py` rebuilds the HTML, Markdown, and feeds from this file.
The `ai-security-digest` GitHub Actions workflow runs both once a day.
