Skill
Open Feed Recsys Reviewer
Review open feed recommendation repositories with source-backed evidence, artifact-readiness checks, and cautious architecture summaries. Use when a user ask...
When to use Open Feed Recsys Reviewer
Choose if
You're auditing claims about open feed recommender repositories (default target xai-org/x-algorithm) and want a disciplined, evidence-backed review — claim support classification (supported / partly_supported / unsupported / not_public_repo), Phoenix artifact-readiness verification, and visible architecture mapping across Home Mixer, retrieval, ranking, Thunder, Grox, candidate pipelines, and ads blending.
Avoid if
You need definitive claims about live production behavior, ranking accuracy, or virality — the README explicitly disclaims these. Also avoid when only private/authenticated sources would settle the question; this skill works strictly from public source plus user-provided files.
Risk Flags
- MEDIUM scope README explicitly limits scope to public source plus user-provided files — no private data or authenticated pages. Phoenix execution is only verified when LFS artifacts are present locally; otherwise execution is unverified.
- MEDIUM scope README states the skill makes no claims about live reach, ranking accuracy, virality prediction, or platform equivalence, and that public code is not guaranteed to contain production weights, thresholds, or all serving rules.
Cost
Type: Free
Distribution
- ClawHub
open-feed-recsys-lab