antigravity-for-claude-code

SUSPICIOUS: the skill’s purpose broadly matches its capabilities, but its footprint is high-risk for a Claude Code skill because it installs a third-party skill, relies on an authenticated external CLI, forwards code/context to that CLI, and combines untrusted web/log ingestion with optional write/apply modes. This looks more like a powerful delegation wrapper than outright malware, but the trust chain and data exposure are significant.