Safety for AI agents. Real-time threat classification to detect malicious content before it causes agents harm.
- Updated install instructions to mention ClawHub and clarify that the skill file can be copied from provided URLs. - Added explicit advice against auto-overwriting skill files; users are directed to review changelogs on Moltbook before updating. - Revised the heartbeat setup instructions to include version checks, changelog review, and periodic self-checks instead of automatic file replacement. - General emphasis on reviewing changes and maintaining user awareness of updates for safety.