A path traversal vulnerability in NagaAgent up to version 5.1.0 allows unauthenticated remote attackers to manipulate filesystem paths and delete or create arbitrary directories.

This is a path traversal flaw (CWE-22) residing in the apiserver/routes/extensions.py component, where attacker-controlled skill names are concatenated directly into filesystem paths without boundary checks. The vulnerability can be triggered by unauthenticated attackers via the skill import and delete API endpoints.

A successful exploit permits unauthorized file system modification, leading to high integrity and potential availability impacts through arbitrary directory creation or recursive deletion. With a CVSS score of 7.3, this flaw presents a significant operational risk, as threat actors can sabotage application data or corrupt the underlying host environment depending on service permissions.

Immediate Action: Restrict API binding and network exposure to trusted local interfaces while awaiting an official vendor patch, as no fixed version is currently available.

Proactive Monitoring: Monitor API access logs for anomalous skill names containing path traversal markers such as dot-dot slash sequences, forward slashes, or backslashes.

Compensating Controls: Implement Web Application Firewall rules or reverse proxy filters to block incoming HTTP requests containing path traversal patterns targeting the /skills endpoints.

Public Exploit Available: Yes, a public proof-of-concept exists via GitHub issue references.

Administrators must treat this vulnerability with high urgency due to the availability of public proof-of-concept code and the lack of an official patch. Network segmentation and strict reverse proxy filtering should be applied immediately to isolate the affected NagaAgent API from untrusted networks until a secure update is released.