substack-publish

SUSPICIOUS. The skill’s stated purpose is plausible, and its shell usage is benign, but its actual trust and data-flow model are not fully coherent: it asks for a high-value browser session token, routes operations through a gateway intermediary, and claims write/publish capabilities that do not align with Substack’s documented official MCP model. No confirmed malware or exploit behavior is present, but the credential scope and opaque backend make this a medium-high security risk skill.