autonomous-dev

SUSPICIOUS: the skill's core behavior is coherent for an autonomous dev workflow, and the supply-chain finding is benign documentation context. The main risk is proportional but significant autonomy: it can execute shell commands, push code, open/update PRs, watch CI, and post comments based on untrusted review content without stepwise confirmation. This is high-impact workflow automation rather than clear malware, but it should be treated as a medium-risk autonomous coding skill.