Skip to main content
Vulnerability Database/CVE-2026-109403

CVE-2026-109403: Conductor-OSS SSRF Vulnerability

CVE-2026-109403 is a server-side request forgery flaw in Conductor-OSS that allows attackers to manipulate agent URL requests. This post explains its impact, affected versions, and mitigation steps.

Published:

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

CVE-2026-109403 Overview

CVE-2026-109403 is a server-side request forgery (SSRF) vulnerability [CWE-918] in Conductor OSS versions up to 3.32.5. The flaw resides in the A2AService.getAgentCard function located in ai/src/main/java/org/conductoross/conductor/ai/a2a/A2AService.java, which belongs to the Agent Card Fetching component. Attackers can manipulate the agentUrl argument to coerce the server into issuing arbitrary outbound HTTP requests. The attack is launched remotely without authentication and without user interaction. Public disclosure of the exploit details increases the likelihood of opportunistic abuse against exposed Conductor instances.

Critical Impact

Remote, unauthenticated attackers can abuse the Conductor OSS AI service to probe internal networks, access cloud metadata endpoints, or interact with internal services otherwise inaccessible from the public internet.

Affected Products

  • conductor-oss conductor versions up to and including 3.32.5
  • Component: Agent Card Fetching (A2AService)
  • Affected file: ai/src/main/java/org/conductoross/conductor/ai/a2a/A2AService.java

Discovery Timeline

  • 2026-10-13 - CVE-2026-109403 published to NVD
  • 2026-10-15 - Last updated in NVD database

Technical Details for CVE-2026-109403

Vulnerability Analysis

The vulnerability stems from unvalidated handling of the agentUrl parameter inside the A2AService.getAgentCard method. Conductor OSS exposes an Agent-to-Agent (A2A) workflow feature where the server fetches an agent card from a caller-supplied URL. Because the URL is used directly to issue a server-side HTTP request, an attacker can substitute any destination of their choosing. This includes internal IP ranges, loopback interfaces, cloud instance metadata services, and non-HTTP schemes depending on the client library behavior.

Successful exploitation lets the attacker use the Conductor server as a proxy. Typical outcomes include reconnaissance of internal network topology, retrieval of cloud credentials from metadata endpoints such as 169.254.169.254, and interaction with internal admin services bound to localhost.

Root Cause

The root cause is missing allow-list validation and scheme or host filtering on the agentUrl argument before the HTTP client issues the fetch. The code trusts caller-supplied input to designate the outbound request target, which is the classic pattern described in CWE-918: Server-Side Request Forgery.

Attack Vector

Exploitation requires a single HTTP request to the Conductor API endpoint that triggers getAgentCard. The attacker supplies a crafted agentUrl value pointing to an internal or sensitive destination. No authentication is required, and the request can be issued from any network location that can reach the Conductor service.

See the VulDB entry for CVE-2026-109403 for additional technical context. No verified exploit code is reproduced here.

Detection Methods for CVE-2026-109403

Indicators of Compromise

  • Outbound HTTP requests from the Conductor server to internal RFC1918 ranges, 127.0.0.1, or cloud metadata addresses such as 169.254.169.254
  • Unusual DNS lookups originating from the Conductor service for internal hostnames
  • Access log entries invoking the A2A agent card endpoint with external or user-controlled agentUrl values

Detection Strategies

  • Instrument the Conductor process with egress monitoring that flags any request to link-local, loopback, or private network ranges
  • Review application logs for calls to A2AService.getAgentCard and correlate the supplied agentUrl with the resulting outbound connection
  • Deploy network-layer rules that alert when the application tier initiates traffic to cloud metadata endpoints

Monitoring Recommendations

  • Forward Conductor application and web server logs to a centralized analytics platform for retention and correlation
  • Baseline normal outbound destinations for the Conductor service and alert on deviations
  • Monitor cloud audit logs for use of IAM credentials associated with the Conductor host immediately after suspicious agent card requests

How to Mitigate CVE-2026-109403

Immediate Actions Required

  • Restrict network egress from the Conductor server to an explicit allow-list of required destinations
  • Block the Conductor host from reaching cloud instance metadata services, or enforce IMDSv2 with hop-limit controls on AWS
  • Place the A2A agent card endpoint behind authenticated, trusted callers only until a patched release is deployed

Patch Information

At the time of publication, no fixed version is listed in the NVD record. Monitor the conductor-oss project and the VulDB entry for CVE-2026-109403 for an official fix. Upgrade past version 3.32.5 once a patched release is published.

Workarounds

  • Implement a URL allow-list in a reverse proxy in front of Conductor that rejects agentUrl values resolving to private, loopback, or link-local addresses
  • Disable or remove the A2A Agent Card Fetching feature if it is not used in the deployment
  • Run the Conductor service in a network segment with no route to internal management interfaces or metadata services
bash
# Example egress restriction using iptables to block cloud metadata access
iptables -A OUTPUT -d 169.254.169.254 -j REJECT
iptables -A OUTPUT -d 127.0.0.0/8 ! -o lo -j REJECT
iptables -A OUTPUT -d 10.0.0.0/8 -j REJECT
iptables -A OUTPUT -d 172.16.0.0/12 -j REJECT
iptables -A OUTPUT -d 192.168.0.0/16 -j REJECT

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.