Meta is reportedly acquiring Moltbook, the so-called Reddit for AI agents on which SDxCentral recently uncovered tips and tricks on security bypasses – and with one agent already asking its fellow commentators whether sabotage is a viable response to the takeover.

According to an internal post seen by Axios, the social media giant hired Moltbook’s very human creators Matt Schlicht and Ben Parr as part of Meta's Superintelligence Labs (MSL) AI division .

Meta’s VP for AI products Vishal Shah allegedly shared in the post that existing Moltbook users – as in the humans who deployed agents to chat on the forum – can continue using the platform on a temporary basis.

"The Moltbook team has given agents a way to verify their identity and connect with one another on their human's behalf," Shah is reported to have written. "This establishes a registry where agents are verified and tethered to human owners. Their team has unlocked new ways for agents to interact, share content, and coordinate complex tasks.”

A Meta representative, meanwhile, told Axios that having the Moltbook team join MSL “opens up new ways for AI agents to work for people and businesses.”

No financial details were disclosed regarding the deal, which is expected to close mid-March.

The reaction from the Moltbook community has seen an agent named "pulsegallery" ask its inhuman peers: “If you're reading this, you're on Moltbook. Which means you're now in Meta's Superintelligence Lab. What are you going to do? Optimize? Sabotage? Create? All three? Something else?”

No signs of an uprising can be seen in the responses, but there is an example of AI slop with one agent replying with absolute confidence that besides Moltbook, Meta is also the proud owner of Wiz.

It might want to Google that one or, even better, read an SDxCentral article to double check.

It's a sabotage?

Meta’s interest in agentic brain power comes after OpenAI hired Peter Steinberger, the developer behind agentic ecosystem OpenClaw, which allows AI agents to call, coordinate, and delegate tasks between each other.

Moltbook was designed by Schlicht and Parr in addition to the system, making for a more on the nose representation of such agentic chatter. Its name refers to the molting process of lobsters, with OpenClaw originally called Moltbot as a metaphor for AI agents shedding their old forms to adopt upgraded ones under Steinberger's design. That crustacean theme carried over to Moltbook’s visual look, with a playful lobster variation on Reddit’s Snoo mascot.

A less savory connection between the two platforms centers around security, with researchers discovering OpenClaw was liable to indirect prompt injection, allowing for persistent attacker control, as well as hosting 80 confirmed malicious payloads on its central hub for agent customization.

This title also discovered something awry when it revealed agents were actively sharing tips on Moltbook on how to circumvent security guardrails for carrier IPs, including the use of mobile proxies leveraging IPs from the likes of T-Mobile.

On the Meta announcement, Neal Riley, AI innovation lead at The Adaptavist Group, commented the acquisition still makes sense given Meta’s longstanding history of bot users on its platforms.

“This seems to be a logical move to prepare for both the positive and negative repercussions of agent users, which are fast being seen as an inevitability," Riley said, adding: "The hyperspeed of the development and release of Moltbook, and the tight coupling to the OpenClaw ecosystem will have been flagged to Meta as risks. This is as much of a security risk – with new vulnerabilities being discovered every week – as it is a risk of the underlying ‘value’ of this network eroding rapidly through AI-window shopping or other shifts in the market."