Daily AI Security Intelligence

Critical flaws raise credential and connected-data exposure risks

SAP disclosed CVE-2026-89111, a CVSS 10 OAuth2 SSRF flaw in Cloud SDK for Python that could expose application client credentials when an attacker controls the tenant subdomain, according to "SAP Discloses Critical CVSS 10 Flaw in Cloud SDK for Python Allowing OAuth2 Credential Theft". GitLab also patched CVE-2026-97229, a CVSS 10 flaw that could enable unauthenticated code execution on affected self-managed servers, potentially affecting repositories and pipelines used for AI development, according to "Critical GitLab Flaw (CVSS 10.0) Could Allow Unauthenticated RCE Across 12 Years of Releases" and "GitLab Patches CVSS 10.0 Flaw That Could Let Unauthenticated Users Run Server Code". SonicWall SMA 1000 exploitation attempts were observed after patching, although successful compromise was not confirmed, according to "Critical SonicWall SMA 1000 Flaw Targeted in 135 Exploitation Attempts After Patch". RealGround analysis: exposed credentials, compromised build infrastructure, or gateway access could provide paths to sensitive AI-connected data, but the available reports do not establish AI-specific compromise. Paperclip's authorization and import weaknesses also illustrate how insuff

2026-10-22 data leakage RealGround analysis
Run AI Security Assessment Talk to AI CISO Download Daily Brief
Top risk today data leakage
Affected industries Healthcare, Fintech, SaaS, SMB, AI startups
Highest severity signal Critical flaws raise credential and connected-data exposure risks
Recommended action Review agent permissions, data access, approval gates, and prompt-injection test coverage.
Relevant RealGround service AI Agent SBOM and LLM Mapping

What Happened

SAP disclosed CVE-2026-89111, a CVSS 10 OAuth2 SSRF flaw in Cloud SDK for Python that could expose application client credentials when an attacker controls the tenant subdomain, according to "SAP Discloses Critical CVSS 10 Flaw in Cloud SDK for Python Allowing OAuth2 Credential Theft". GitLab also patched CVE-2026-97229, a CVSS 10 flaw that could enable unauthenticated code execution on affected self-managed servers, potentially affecting repositories and pipelines used for AI development, according to "Critical GitLab Flaw (CVSS 10.0) Could Allow Unauthenticated RCE Across 12 Years of Releases" and "GitLab Patches CVSS 10.0 Flaw That Could Let Unauthenticated Users Run Server Code". SonicWall SMA 1000 exploitation attempts were observed after patching, although successful compromise was not confirmed, according to "Critical SonicWall SMA 1000 Flaw Targeted in 135 Exploitation Attempts After Patch". RealGround analysis: exposed credentials, compromised build infrastructure, or gateway access could provide paths to sensitive AI-connected data, but the available reports do not establish AI-specific compromise. Paperclip's authorization and import weaknesses also illustrate how insuff

Why This Matters

AI systems increasingly connect natural-language decisions to SaaS integrations, internal data, memory stores, API calls, and production workflows. A signal that appears narrow in a vendor report can become broader business risk when it intersects with autonomous tools or sensitive context.

Healthcare Fintech SaaS SMB AI startups

RealGround Analysis

This trend increases exposure to indirect prompt injection, unauthorized tool execution, sensitive data disclosure, and weak human approval workflows for organizations deploying LLM agents or AI-enabled automation.

Recommended Actions

  • Inventory SAP Cloud SDK for Python, GitLab, SonicWall SMA 1000, Paperclip, and AI-connected integrations in scope.
  • Apply the relevant vendor fixes, including fixed SAP SDK versions, GitLab 19.4.2, 19.3.5, or 19.2.8 where applicable, and SonicWall hotfixes.
  • Validate tenant identifiers and reject attacker-controlled or untrusted subdomains in OAuth2 and BTP integration workflows.
  • Rotate potentially exposed OAuth2 client secrets and review authentication, access, and downstream data-use logs for anomalous activity.
  • Review GitLab CI/CD configurations, runner permissions, deployment credentials, and recent artifacts for unauthorized changes.
  • Strengthen agent-control-plane authorization, self-registration restrictions, configuration-import validation, and privilege-boundary testing.
  • Restrict agent permissions with least-privilege tool scopes.
  • Add human approval workflows for state-changing actions.
  • Review SaaS integrations, memory persistence, and data access paths.
  • Test prompt injection and indirect prompt injection scenarios before production rollout.

Relevant RealGround Service

AI Agent SBOM and LLM Mapping Schedule security review AI Security Readiness Assessment Schedule security review AI Agent Business Logic Audit Schedule security review AI Red Teaming Schedule security review AI CISO Advisory Schedule security review

Sources

SAP Discloses Critical CVSS 10 Flaw in Cloud SDK for Python Allowing OAuth2 Credential Theft Critical GitLab Flaw (CVSS 10.0) Could Allow Unauthenticated RCE Across 12 Years of Releases GitLab Patches CVSS 10.0 Flaw That Could Let Unauthenticated Users Run Server Code Critical SonicWall SMA 1000 Flaw Targeted in 135 Exploitation Attempts After Patch Critical Paperclip Flaw Allowed Admin Access, Code Execution
Talk to AI CISO