The vulnerability resides in the shell.exec function within the hermes-agent server.py module. A malicious actor can manipulate the execution flow to bypass the built‑in protection mechanism. This failure permits the execution of arbitrary shell commands on the system running the agent, which is classified as CWE‑693. The description states the attack can be launched remotely, so attackers do not need local access to exploit the flaw.

The affected product is NousResearch hermes-agent, specifically all releases up to and including version 0.15.2. The component that is vulnerable exposes a server interface, server.py, which provides remote shell execution capabilities. No other versions or related products are listed as vulnerable by the CNA.

The CVSS score of 5.3 places the issue in the medium severity range. The EPSS score of less than 1% indicates a low probability of widespread exploitation, and the vulnerability is not listed in the CISA KEV catalog. Based on the description, the likely attack vector is remote network traffic targeting the hermes‑agent server.py endpoint, as the vulnerable function enables execution of arbitrary commands without proper protection. Attackers would send crafted requests to the exposed endpoint to override the protection mechanism and run commands, which could compromise confidentiality, integrity, and availability of the host. While the public supply of the exploit increases the potential risk, the low EPSS score suggests that exploitation remains relatively uncommon at this time.