agentleFS
getedgehqgetedgehq/skills/broll-picker/SKILL.md

broll-picker

Skill5 starsChanged 11 days agoApache-2.0View on GitHub

Find B-roll in the user's iCloud Photos, meaning clips and photos of places and scenes without the user or close-up people. Filter by length, place and scene labels (city, sunset, office, food), rank by Apple's aesthetic scores, keep variety across moments, review visually and download originals. Use for video edits, reels, and post backgrounds.

The file

Files
SKILL.md

B-roll picker

Needs icloud-photos-bridge set up (its setup.sh saves where its scripts live).

PY=$(cat ~/.config/icloud-photos/python 2>/dev/null); B=$(cat ~/.config/icloud-photos/bridge_dir 2>/dev/null)
[ -x "$PY" ] && [ -d "$B" ] || echo "icloud-photos-bridge is not set up: run its scripts/setup.sh first"

scripts/pick.py below is in this skill's own folder.

Steps

  1. Catalog: $PY $B/catalog.py --out catalog.jsonl --since <date> --kind all
  2. Rank:
    $PY scripts/pick.py catalog.jsonl --owner $($PY $B/owner.py get) --kind video --top 48 \
        [--min-sec 3 --max-sec 30] [--label city --label sunset] [--place "Lisbon"]
    • Leaves out selfies, anything with the owner in it, and by default anything with faces or people labels. --people small allows small background people (street scenes).
    • At most one pick per 10 minutes (--window-min), so you get different scenes.
    • --kind photo|all for stills. Labels are Apple's scene labels; look at a few catalog lines to see the vocabulary.
  3. Look before you pick. $PY $B/montage.py broll_picks.json m.jpg --n 48 (videos show their poster frame). Drop: tilted horizons, blown-out sky, screens or documents showing private info, license plates up close, anything where a person ends up as the subject. Keep a mix: wide establishing shots, details, motion.
  4. Show the user $PY $B/sheet.py final.json broll.html --title "B-roll" with duration and labels per item.
  5. Download originals: $PY $B/fetch.py ./broll-<date> --from uuids.txt. Videos can be large; check free disk first.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.