atomic-agents-stack: Dashboard HTTP server path traversal allows arbitrary file read
The atomic-agents-stack dashboard HTTP server has a path traversal vulnerability in its file serving component. It constructs filesystem paths directly from HTTP request paths without proper containment checks, allowing attackers to read arbitrary files outside the intended directory. By default, the server binds to loopback, but it can be configured to bind to all interfaces, exposing the vulnerability to the LAN. This flaw affects all versions prior to 1.1.0. A fix is available that enforces safe path resolution and restricts host binding.
AI Analysis
Technical Summary
The vulnerability exists in the atomic-agents-stack dashboard HTTP server (dashboard/serve.py), specifically in the DashboardHandler.do_GET and _serve_file methods. The server builds filesystem paths from the request path without routing through the safe path resolution function (_io.safe_resolve_under), allowing literal '../' segments to survive URL parsing and path joining. This enables arbitrary file read outside the intended agents_root directory. The default server bind is loopback, but the --host flag can be set by operators to bind to 0.0.0.0, exposing the vulnerability to the local network. Even on loopback, DNS rebinding or SSRF can be used to exploit the flaw. The vulnerability affects all versions before 1.1.0. The recommended fix is to route all served paths through _io.safe_resolve_under and reject path traversal attempts early, and optionally restrict non-loopback binding unless explicitly authorized.
Potential Impact
An attacker can read arbitrary files on the server filesystem outside the intended directory, potentially exposing sensitive information. The vulnerability is exploitable remotely if the server is bound to a non-loopback interface or via DNS rebinding or SSRF attacks even when bound to loopback. This can lead to information disclosure without requiring authentication or user interaction.
Mitigation Recommendations
A patch is available that routes all served paths through the safe path resolution function (_io.safe_resolve_under) and rejects requests with path traversal attempts. Operators should upgrade to version 1.1.0 or later. Additionally, restricting the server to bind only to loopback interfaces by default and requiring explicit authorization for non-loopback binding reduces exposure. Rejecting '..' and separator characters early in request paths further mitigates the risk.
atomic-agents-stack: Dashboard HTTP server path traversal allows arbitrary file read
Description
The atomic-agents-stack dashboard HTTP server has a path traversal vulnerability in its file serving component. It constructs filesystem paths directly from HTTP request paths without proper containment checks, allowing attackers to read arbitrary files outside the intended directory. By default, the server binds to loopback, but it can be configured to bind to all interfaces, exposing the vulnerability to the LAN. This flaw affects all versions prior to 1.1.0. A fix is available that enforces safe path resolution and restricts host binding.
CVSS v4.0
Affected software
Run on your own infrastructure? Check whether these packages are installed with threat-finder — our free open-source scanner.
Weaknesses
AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
The vulnerability exists in the atomic-agents-stack dashboard HTTP server (dashboard/serve.py), specifically in the DashboardHandler.do_GET and _serve_file methods. The server builds filesystem paths from the request path without routing through the safe path resolution function (_io.safe_resolve_under), allowing literal '../' segments to survive URL parsing and path joining. This enables arbitrary file read outside the intended agents_root directory. The default server bind is loopback, but the --host flag can be set by operators to bind to 0.0.0.0, exposing the vulnerability to the local network. Even on loopback, DNS rebinding or SSRF can be used to exploit the flaw. The vulnerability affects all versions before 1.1.0. The recommended fix is to route all served paths through _io.safe_resolve_under and reject path traversal attempts early, and optionally restrict non-loopback binding unless explicitly authorized.
Potential Impact
An attacker can read arbitrary files on the server filesystem outside the intended directory, potentially exposing sensitive information. The vulnerability is exploitable remotely if the server is bound to a non-loopback interface or via DNS rebinding or SSRF attacks even when bound to loopback. This can lead to information disclosure without requiring authentication or user interaction.
Mitigation Recommendations
A patch is available that routes all served paths through the safe path resolution function (_io.safe_resolve_under) and rejects requests with path traversal attempts. Operators should upgrade to version 1.1.0 or later. Additionally, restricting the server to bind only to loopback interfaces by default and requiring explicit authorization for non-loopback binding reduces exposure. Rejecting '..' and separator characters early in request paths further mitigates the risk.
Technical Details
- Gcve Source
- db.gcve.eu
- Osv Id
- GHSA-rm43-82j9-r4mj
- Osv Schema Version
- 1.4.0
- Aliases
- []
- Ecosystems
- ["PyPI"]
- Database Specific Severity
- HIGH
- Cvss Version
- 4.0
Threat ID: 6a7e036fbf8831d5398f9371
Added to database: 08/13/2026, 17:48:31 UTC
Last enriched: 08/13/2026, 18:10:20 UTC
Last updated: 08/13/2026, 18:10:20 UTC
Views: 2
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
External Links
Need more coverage?
Upgrade to Pro Console for AI refresh and higher limits.
For incident response and remediation, OffSeq services can help resolve threats faster.
Latest Threats
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.