Yummy Shared

This skill looks like a legitimate yummycli helper, but there are a few red flags you should address before enabling it widely: - Confirm why the skill needs GEMINI_API_KEY as a required primary env var. If the skill only checks whether credentials exist, it should be able to run defensively without requiring the secret at load time. If the skill will perform authenticated operations, ensure you trust the maintainer and scope the token with minimum permissions. - Ask the publisher to justify always: true. That setting forces the skill into every agent run and expands the risk of any sensitive env var being accessible. Prefer enabling the skill on demand or only when yummycli is being used. - Clarify the install behavior. The SKILL.md includes an npm install block for @yummysource/yummycli, but the registry metadata stated no install spec — ask whether the agent will auto-install this package. If it will, verify the npm package exists, inspect its source, and confirm the package maintainer/trustworthiness before allowing installation. - Because the skill has no homepage/source listed, request a repository or package link so you can inspect code and provenance. - If you decide to allow it: limit scope by providing a scoped GEMINI_API_KEY (least privilege), avoid making the skill always-on, and enable monitoring (audit logs, token rotation) in case of unexpected behavior. If the publisher cannot explain the always: true setting, the install inconsistency, or provide a verifiable source, treat the skill as high-risk and avoid enabling it.