Develop topic

Turn a candidate topic into an approved content contract. The work is ready for writing only when the argument, evidence, and boundaries no longer need to be invented downstream.

Scope

Start from one candidate topic or a small, naturally bounded set. When several nearby directions remain, compare them lightly for audience value, novelty, verifiability, demonstrability, author fit, and overclaiming risk; ask the user to choose before deep research. Open-ended trend scanning and maintaining an idea backlog stay outside this skill.

Evidence workspace

Create a content/ directory in the video project:

content/
|-- RESEARCH.md
|-- TOPIC.md
`-- experiments/