content-atomizer
Content Atomizer
Turn one finished piece into the few platform-native assets it actually deserves, and nothing else. The judgment this skill carries: platform selection and native depth are the work; volume is the failure mode. Rephrased-spray content is a strategy for being ignored at scale; platform-specific evidence lives in references/platform-facts.md. This skill is a floor, not a ceiling. If the content in front of you deserves a better path than the one below, take it and say why in one line.
Context
Only if ./brand/, ./brands/ or MARKETING.md exists (nothing to load otherwise), run:
python3 ${CLAUDE_SKILL_DIR}/../start-here/scripts/brand_context.py --for content-atomizer
(If the script or python3 is unavailable, read the files named in reads: directly, section-anchored where specified. If none exist, proceed standalone and ask only for what the task strictly requires.)
What each input does here:
- voice-profile#platform-adaptations: the per-platform voice table. It is the ONLY source of voice adjustment in this skill; never invent a parallel adaptation table. No profile means ask one question ("how formal, and any words you'd never use?") and write in a clean neutral register.
- learnings: platform performance priors. Apply them. Mention one only when it changed the work, as a single line in the note ("Your carousels outperformed text posts 3/3 in May, so LinkedIn gets one").
- algo-cache: per-brand dated platform facts, if /performance-review or a prior refresh has written any. Where it conflicts with this skill's reference file, the fresher
last-verifiedwins.
Freshness gate (run before generating): load references/platform-facts.md. If today is past last-verified + ttl-days, refresh per the instruction in its header, update the file, and say so in one line in the note. Never generate against stale algorithm facts silently.