Raw vector
CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:HSummary
CVE-2025-58768 is a critical-severity Code Injection (CWE-94) vulnerability in Thinkinai Deepchat. Its CVSS base score is 9.6 (Critical).
Operationally, exploitation aligns with the MITRE ATT&CK technique Browser Session Hijacking (T1185); ranked at the 46th percentile by exploit likelihood (below the median); it is not currently listed in the CISA KEV catalog; a public proof-of-concept is referenced.
This vulnerability is AI-related — categorised as LLM Application Platforms; in the Other ATLAS/OWASP Terms risk domain.
The strongest mitigations our analysis identified map to SA-11 (Developer Testing and Evaluation) and SI-10 (Information Input Validation) — see the control section below for these in your framework.
Deeper analysis AI-assisted summary
Synthesised by an AI model from the NVD description and linked references — a reading aid, not an authoritative source.
CVE-2025-58768 is a high-severity vulnerability (CVSS 9.6) affecting DeepChat, an artificial intelligence-powered smart assistant, in versions prior to 0.3.5. The issue lies in the Mermaid chart rendering component, which unsafely uses innerHTML to set user-supplied content. This enables cross-site scripting (XSS, CWE-79) that triggers an exploit chain, ultimately leading to arbitrary code injection (CWE-94) and command execution via exposed IPC. The flaw stems from an incomplete fix for a prior XSS vulnerability documented in GHSA-hqr4-4gfc-5p2j.
The vulnerability is exploitable remotely (AV:N) with low attack complexity (AC:L), requiring no privileges (PR:N) but user interaction (UI:R), and it changes scope (S:C) to achieve high impacts on confidentiality, integrity, and availability (C:H/I:H/A:H). An attacker can craft malicious Mermaid content that, when rendered by a user in DeepChat, executes arbitrary JavaScript via XSS and escalates to arbitrary command execution through the exposed IPC interface.
The GitHub security advisory (GHSA-f7q5-vc93-wp6j) recommends updating to DeepChat version 0.3.5, which includes an updated fix for the innerHTML handling in the Mermaid renderer. No additional workarounds are specified.
This vulnerability is notable in the context of AI applications, as DeepChat's smart assistant functionality integrates AI with user-generated content rendering, exposing a chain from XSS to system-level command execution. No public reports of real-world exploitation are available as of the CVE publication on 2025-09-09.
OWASP Top 10 for Web (2025)
EU & UK References
- 🇪🇺 ENISA EUVD: EUVD-2025-27502
Vulnerability Data
DeepChat is a smart assistant uses artificial intelligence. Prior to version 0.3.5, in the Mermaid chart rendering component, there is a risky operation of directly using `innerHTML` to set user content. Therefore, any malicious content rendered via Mermaid will directly…
more
trigger the exploit chain, leading to command execution. This vulnerability is primarily caused by a failure to fully address the existing XSS issue in the project, leading to another exploit chain. The exploit chain is consistent with the report GHSA-hqr4-4gfc-5p2j, executing arbitrary JavaScript code via XSS and arbitrary commands via exposed IPC. Version 0.3.5 contains an updated fix.
- CWE(s)
AI Security AnalysisAI
- AI Category
- LLM Application Platforms
- Risk Domain
- Other ATLAS/OWASP Terms
- OWASP Top 10 for LLMs 2025
- None mapped
- Classification Reason
- Matched keywords: artificial intelligence
Related Threats
MITRE ATT&CK Enterprise Techniques
CVEs Like This One
Affected Assets
Mitigating Controls
Control response
—
—
—
V1.1.2V1.3.2V1.3.1
Mitigating Controls (NIST 800-53 r5) AI
Developer testing and evaluation finds code paths that accept and execute externally influenced strings.
Input validation directly stops untrusted data from being used to construct executable code without neutralization.
Output filtering can catch or sanitize unneutralized script content before it is served to users.
Least privilege limits the damage an injected code fragment can perform once executed.
Requiring documented secure development standards and tools enforces use of safe code-generation APIs and escaping.
Mitigating Controls (NIST CSF 2.0) AI
Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→CSF cross-walk (authority under review) — links open the control.
PR.PS-06's SDLC practices directly target injection flaws via secure coding and testing (mostly), yet as a single broad outcome it leaves many code-generation specifics unaddressed (partial).
Patching and EOL replacement can remediate known XSS instances in libraries or frameworks (partial) but do nothing to enforce input neutralization in application code (none).
PR.DS-10 protects runtime data confidentiality/integrity but has no bearing on neutralizing externally influenced input during code generation, so neither direction shows any preventive effect.
Mitigating Controls (ISO/IEC 27001:2022 Annex A) AI
Derived directly from the weakness types (CWEs) cited in the NVD entry via our AI-authored CWE→ISO cross-walk (authority under review) — links open the control.
Secure-coding testing and automated code-analysis tools are applied to detect improper neutralization of script-related content during web-page generation.
Knowledge exchange on emerging attack techniques and patches reduces the likelihood that cross-site scripting flaws remain unaddressed in deployed applications.
Operational indicators of compromise for web-application attacks can be incorporated into WAF or input-filtering rules, lowering the likelihood that unsanitized data reaches the browser.
Requiring language-specific secure-coding standards and automated scanning during the SDLC catches missing output encoding or improper neutralization of untrusted data before the software reaches production.
Banning unapproved code samples and unauthenticated web services, combined with secure-coding standards and SAST, prevents the dynamic generation or inclusion of attacker-supplied code.
Controls that restrict unauthorized or malicious code from being introduced via external networks or removable media limit opportunities for an attacker to inject and execute arbitrary code.