CVE-2026-44118
Published: 06 May 2026
Description
OpenClaw before 2026.4.22 derives loopback MCP owner context from spoofable server-issued bearer tokens in request headers. Non-owner loopback clients can present themselves as owner to bypass owner-gated operations by manipulating the sender-owner header metadata.
Likely Mitigating ControlsAI
Per-CVE control mapping for this CVE has not run yet; the list below is derived from the weakness types (CWEs) cited in the NVD entry.
Reveals spoofed logon attempts through unexpected previous logon timestamps upon legitimate login.
Training specifically addresses recognizing spoofed communications and phishing that enable authentication bypass.
Requiring verifiable identity evidence at appropriate assurance levels makes it substantially harder for attackers to successfully spoof or impersonate users to obtain accounts.
Unique device authentication makes successful spoofing of device identity substantially more difficult to achieve.
Unique identification of non-organizational users reduces the feasibility of authentication bypass by spoofing.
Unique identification and authentication of services before communications makes spoofing of service identities substantially harder.
Isolated trusted path ensures the user interacts only with genuine system components, preventing spoofing of authentication interfaces or prompts.
Directly counters DNS response spoofing by requiring cryptographic origin authentication artifacts from the authoritative source.
Security SummaryAI
Automated synthesis unavailable for this CVE.
Details
- CWE(s)
AI Security AnalysisAI
- AI Category
- AI Agent Protocols and Integrations
- Risk Domain
- Protocol-Specific Risks
- OWASP Top 10 for LLMs 2025
- None mapped
- MITRE ATLAS Techniques
- None mapped
- Classification Reason
- Matched keywords: mcp