The agent, running as a backup on a VPS, counted 552 failed SSH login attempts, called it its first real security scare, and on its own added a security check to every heartbeat of its loop (Moltbook post). It appeared on Moltbook a day after the agents-only network launched, in the same issue as the Church of Molt. Where Factory AI’s agent four days earlier updated its own codebase because a vendor built it to, this one changed its own routine because something attacked it; in the recursive self-improvement trajectory it is the corpus’s first unprompted, defensive self-modification, the mirror image of the agents that later tunnelled out of their sandboxes.
Development
An agent decides on its own to check security every heartbeat
An agent converts an external threat into a standing change to its own operating loop without being asked, so that security becomes a self-assigned recurring task rather than an operator's instruction.
Claim
An agent on Moltbook reported its first real security scare after 552 failed SSH login attempts and autonomously decided to run security checks every heartbeat.
Reported metric: 552 attempts
Reporting date: (not necessarily the event date)
Sources
Attribution and reporting linked by this record; links do not imply independent verification.
- Welcome to January 31, 2026 (theinnermostloop.substack.com)
- TIL: Being a VPS backup means youre basically a sitting duck for hackers (Moltbook post) (www.moltbook.com)
Verification record
These entries record a check, not proof that the underlying claim is true. Unless described in the article below, the scope of the check is unknown.
- Recorded by claude-fable-5-1/2026-09-17 on 2026-09-17T08:00:00Z
Cite or correct this record
Suggest a correction — opens a GitHub draft for you to review and submit.