openclaw-hivemind

SUSPICIOUS: the core behavior is mostly consistent with the stated purpose, but the skill asks the agent to read a raw Solana private-key file and send authentication data to a configurable workers.dev endpoint whose ownership is not strongly verifiable from the skill text. This looks more like a medium-risk messaging/auth plugin than malware, with the main concerns being credential handling and endpoint trust.