Vulnerabilidade XSS no Mermaid leva à Execução Remota de Código

The impact of CVE-2025-67744 is severe. An attacker can exploit this vulnerability to execute arbitrary code on the affected system. This is achieved by crafting malicious Mermaid diagrams that, when rendered, inject and execute JavaScript code. The exposed Electron IPC renderer allows this injected code to interact with the underlying system, effectively granting the attacker remote control. This could lead to data theft, system compromise, and potential lateral movement within the network. The combination of XSS and IPC exposure creates a highly dangerous attack vector, similar in impact to vulnerabilities that bypass security sandboxes.

Organizations utilizing DeepChat for AI agent development and deployment are at risk, particularly those relying on older versions (≤ 0.5.3). Shared hosting environments where DeepChat is deployed alongside other applications are also at increased risk, as a successful exploit could potentially compromise the entire host.

• javascript / web: Inspect DeepChat's JavaScript code for instances of eval() or new Function() calls related to Mermaid rendering.

// Example: Search for calls to eval() within Mermaid rendering functions
console.log(code.toString().match(/eval\s*\(/g) || []);

• javascript / web: Monitor network traffic for unusual JavaScript payloads being sent to the DeepChat application. • javascript / web: Examine the Electron IPC channel for unexpected messages or commands being executed. • generic web: Review DeepChat's access and error logs for any signs of XSS attempts or unusual activity related to Mermaid diagrams.

1. 2025-12-16Disclosure

   disclosure

1. Reservado2025-12-11
2. Publicada2025-12-16
3. EPSS atualizado2026-03-23

The primary mitigation for CVE-2025-67744 is to immediately upgrade DeepChat to version 0.5.3 or later, which contains the necessary patch. If upgrading is not immediately feasible, consider implementing temporary workarounds. While a direct WAF rule is unlikely to be effective against this type of XSS, carefully scrutinizing input to the Mermaid renderer for suspicious characters or patterns could offer limited protection. Review and restrict access to the Electron IPC interface to prevent unauthorized interactions. After upgrading, confirm the fix by attempting to render a known malicious Mermaid diagram and verifying that no code execution occurs.