The source package is too sparse to write a proper article without fabricating content — which the rules explicitly forbid.
Here’s what the package actually contains:
- 1 YouTube video — title and URL only
- Summary: empty string
'' - Comments: 0
- Competitor data: none
- Opinions/tool URLs: none
The video content was never extracted, so there’s no actual information to curate. Writing a 1500–2000 word article from just a video title would mean inventing everything — prices, tool names, features, verdicts — which violates the core rule “Use ONLY information from the source package.”
What you can do to fix this:
- Re-run the research engine with the YouTube transcript extraction enabled for
https://www.youtube.com/watch?v=RKbmqSRc0z0— thesummaryfield needs to be populated - Add more sources — at least 3–5 articles, Reddit threads, or additional YouTube videos on the topic
- Pass the enriched package back and I’ll write the article immediately
If your research engine has a transcript/summary fetcher, the YAML should show something in the summary: field and ideally a num_comments > 0. Right now it looks like the fetch silently failed or was skipped.