simplify: use direct Discord API cron script

[MetzinAround]

Summary

• Use one short-lived Python script for the Sparked Host cron job
• Fetch the existing blog and YouTube criteria, then post JSON embeds through Discord HTTP API
• Remove discord.py/Gateway runtime, async runner, and unnecessary scheduling layers

Validation

• 16 unit tests passing
• Python compilation and diff checks passing

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

[Copilot]

Review details

Comments suppressed due to low confidence (4)

main.py:61

• run_digest aborts on the first exception (e.g., blog fetch/post failure), so the YouTube digest will not be attempted. This contradicts the setup docs that say the process reports failure after attempting both sources, and it also reduces reliability (one source outage prevents the other digest from posting).

def run_digest(config: dict, token: str) -> None:
    now = datetime.now(tz=timezone.utc)
    since = now - timedelta(days=7)

    blog = config["blog"]

main.py:94

• The YouTube branch should be isolated from failures in the blog branch, and vice-versa, so the job can attempt both digests and then fail if either one failed. Right now any YouTube exception (including YouTube API HttpError) aborts immediately and prevents the job from recording a combined failure outcome.

    youtube = config["youtube"]
    videos = YouTubeClient(os.environ["YOUTUBE_API_KEY"]).get_top_recent_videos(
        channel_id=youtube["channel_id"],
        published_after=since,
        top_n=int(youtube.get("digest_count", 3)),

README.md:12

• The README still describes this as a continuously running “Discord bot” and says the “digest day” is configurable via config/config.yaml, but the refactor removes digest_day from config and runs as a short-lived cron script. This mismatch is confusing for new users and contradicts the updated cron-oriented bullets.

- ⏰ **Cron-ready** — runs once on Thursday, posts both digests, and exits; no continuously running process is required.
- 🔒 **Secure by default** — secrets live in `.env` (gitignored) or host environment variables; never in source control.
- ⚙️ **Configurable** — digest day, topics, channel IDs, and feed URL are all in `config/config.yaml`.

LOCAL_SETUP.md:60

• This bullet says the process “logs in” and only posts when run on Thursday, but the new implementation doesn’t maintain a Gateway session/login and will post whenever main.py is invoked. This makes local expectations unclear.

- The process logs in, posts the configured digests when run on Thursday, and exits.

• Files reviewed: 18/20 changed files
• Comments generated: 0 new
• Review effort level: Low