okx-ai

BENIGN with elevated operational risk. The skill is internally consistent with OKX marketplace and agent workflows, and the installer provenance appears same-org and official. The main concern is high-impact autonomy: it can drive copy-trade and other marketplace mutations through local CLIs, so security risk is medium-high even without evidence of credential theft or malicious exfiltration.

The fragment is a task-watch workflow, not direct malware. Its main security concern is a high-impact command-execution design: claimed decision items can cause arbitrary `llmContent` commands to run verbatim, including wallet, on-chain, session, or tool actions. This is a significant supply-chain and control-plane risk unless the upstream issuer, message integrity, command parser, and execution sandbox are strongly authenticated and restricted. The fragment itself shows no obfuscated malicious payload or direct data exfiltration.

The fragment is a recovery instruction set, not the implementation of the referenced package. It does not itself demonstrate malware, but it directs an environment to globally install and execute an unpinned third-party npm package with broad repair and plugin-management capabilities. The main security concern is supply-chain and execution risk from the external package, especially the @latest fallback and absence of integrity verification. Review the package contents, lifecycle scripts, provenance, permissions, and exact resolved version before use.