Mistral AI configuration hardening security checks
Vendor-recommended secure defaults, patch levels and housekeeping settings that drift as tenants grow and admins change.
On Mistral AI, Black Cat runs 3 checks in this area on every scan. Each one below lists its severity, how to fix it, and the compliance controls it satisfies where a control applies. See what access the Mistral AI connector needs.
Checks (3)
severity: medium Agent Uses Code Interpreter fix difficulty: easy #
Review agents granted the code-interpreter (arbitrary code) tool
- Confirm the agent needs code execution
- Remove the code_interpreter tool if not required
Satisfies: NIS2 Directive NIS2-21.e.4 DORA (SaaS Security) DORA-9.10
severity: high Agent References Org-Shared Library fix difficulty: medium #
Re-scope libraries an agent exposes that are shared org-wide
- Identify the org-shared library the agent references
- Restrict the library's sharing or detach it from the agent
Satisfies: NIS2 Directive NIS2-21.e.4 DORA (SaaS Security) DORA-9.10
severity: low Agent Deployed Without Guardrails fix difficulty: medium #
Add guardrails to chat-deployed agents
- Open the agent's configuration
- Configure guardrails before exposing it in chat
Satisfies: NIS2 Directive NIS2-21.e.4 DORA (SaaS Security) DORA-9.10