According to The Register, security concerns continue to surround moltbot ai, the open-source agentic AI personal assistant formerly known as Clawdbot. While the tool has seen rapid adoption among developers, its deep access to personal and professional data has raised serious questions about the risks of deploying such systems without strong security controls.

Agentic capabilities raise data access risks

Moltbot AI can be controlled through messaging platforms such as WhatsApp and Telegram and is designed to handle tasks including email management, calendar scheduling, and reservation bookings. To perform these functions, moltbot ai requires access to sensitive user credentials, including email accounts, encrypted messengers, phone numbers, and even bank-related information. Security experts warn that granting this level of access to an internet-exposed system significantly increases the potential attack surface.

Misconfigurations and supply chain exposure

Researchers have identified hundreds of instances where moltbot ai deployments were exposed online due to misconfigurations. In addition, a supply chain exploit targeting ClawdHub, the skills library used by Moltbot, demonstrated how attackers could execute commands remotely and exfiltrate sensitive data such as SSH keys and AWS credentials. These incidents highlight how weaknesses in the surrounding ecosystem can be leveraged to compromise users.

Plaintext storage and malware threats

Another key concern is how secrets shared with moltbot ai are stored. According to the report, sensitive information is saved in plaintext on local filesystems, leaving it vulnerable to infostealer malware. This design choice further amplifies the risk, particularly for users running the assistant on personal machines without advanced security protections.

A wider challenge for AI agents

The issues facing moltbot ai reflect a broader challenge in the rapid adoption of agentic AI tools. Experts cited by The Register point to a growing gap between user enthusiasm and the technical expertise required to operate these systems securely. Because agentic AI often bypasses traditional security boundaries, specialists argue that existing cybersecurity models need to be reassessed. Without measures such as encryption-at-rest, containerization, strict monitoring, and least-privilege access, the local-first AI trend could become an attractive target for cybercriminals.

The ongoing security concerns around moltbot ai underscore the risks inherent in powerful agentic AI tools that handle sensitive data. As regulators and industry leaders urge caution, the case of moltbot ai highlights the need for stronger security practices to protect both personal and corporate information.