- Rewrote broad_scan.py with:
- Caching layer (.agents/cache/research/) with 6-hour TTL
- Proper Firefox User-Agent for Reddit RSS
- Fallback to old.reddit.com RSS when www is blocked
- Exponential backoff retry logic
- Cache fallback when fresh fetch fails
- Seeded initial cache with all 3 subreddits (weightroom,
advancedfitness, StrongerByScience)
- Fixed save path to logs/research/
- Added .agents/cache/ to .gitignore
First research scan now has real data from all sources