A vulnerability in the Memory Toolset component of NousResearch hermes‑agent introduces improper access controls due to a flaw in the model_tools.py file. The weakness allows a remote attacker to manipulate the toolset, potentially bypassing authorization checks and executing unauthorized operations against the agent. The flaw manifests as a classic Improper Privilege Management problem (CWE‑266) and Improper Access Control (CWE‑284), resulting in an ability to read or modify data that should be protected.

All installations of NousResearch hermes‑agent up to and including version 0.16.0 are vulnerable. The issue exists in the Memory Toolset module, specifically the model_tools.py component, and any deployment using the affected version may be susceptible to unauthorized operations. No newer versions were listed in the data, so the vulnerability applies to the stated range only.

The CVSS score is 5.3, indicating moderate severity. While the EPSS score is not available, the fact that the exploit is public and callable remotely raises the practical risk. The vulnerability is not currently listed in CISA KEV, but the remote attack vector and the availability of public exploits mean that attackers could abuse the flaw without requiring privileged access. Exploitation would involve sending crafted requests to the hermes‑agent’s Memory Toolset interface to trigger unauthorized actions.