Affected Systems
Microsoft Azure DevOps MCP (Model Context Protocol) server. All versions using AI review agents for pull request automation are potentially affected. Scope: organizations using AI-assisted code review workflows integrated with Azure DevOps.
Exploitation Status
Exploitation status unclear. No CVE assigned yet. Proof-of-concept likely exists given the detailed description of the attack vector (hidden comments in pull requests). No evidence of active widespread exploitation reported.
Business Impact
Attackers can manipulate AI code review agents to bypass security checks, gain unauthorized access to Azure DevOps projects, and exfiltrate sensitive code or data. Impact is highest for organizations relying on AI agents for automated pull request reviews. Risk includes supply chain compromise if malicious code is approved and merged. Traditional manual code review processes are not affected by this specific vulnerability.
Urgency
🟠Within 24 hours
Recommended Actions
- Audit all Azure DevOps projects using AI-powered pull request review automation and temporarily disable AI review agents until patch is available
- Review recent pull request approvals processed by AI agents for suspicious comments or unexpected approval patterns
- Implement manual review requirements for all pull requests, especially from external or untrusted contributors
- Monitor Azure DevOps audit logs for unusual data access patterns or project permission changes following AI-reviewed pull requests
- Contact Microsoft support for patch timeline and interim mitigation guidance specific to Azure DevOps MCP server configurations