Skip to content

Ean Schuessler / mo-mcp

Go to a project
Toggle navigation
Toggle navigation pinning
2e5c8995 by Ean Schuessler
Options

Fix security model: plugin uses ADMIN for discovery, enforces user permissions -… 

… deterministic and secure
1 parent aeecc016
Inline Side-by-side
Showing 1 changed file with 12 additions and 12 deletions
...@@ -37,23 +37,23 @@ Foundation for autonomous business operations (ECA/SECA systems). ...@@ -37,23 +37,23 @@ Foundation for autonomous business operations (ECA/SECA systems).
37 37
38 **⚠️ CONTAINERS & SECURITY REQUIRED ⚠️** 38 **⚠️ CONTAINERS & SECURITY REQUIRED ⚠️**
39 39
40 ## 🛡️ **Security: AI User Avatars** 40 ## 🛡️ **Security: Deterministic by Design**
41 41
42 AI agents authenticate as **Moqui users** with **role-based permissions** - same security as human employees. 42 **Plugin uses ADMIN context for discovery, enforces user permissions for access.**
43 43
44 ### **Safe Privilege Escalation** 44 ### **How It Works**
45 - **Start Limited**: AI begins with basic permissions (read-only catalog access) 45 - **Discovery Phase**: Plugin uses ADMIN context to find all available screens
46 - **Earn Trust**: Proven performance triggers Moqui status transitions 46 - **Permission Check**: User permissions enforced before any screen access
47 - **Role Progression**: AI "promotes" from trainee → specialist → manager 47 - **No Escalation**: Users can only access screens they're authorized for
48 - **Business Rules**: All actions constrained by Moqui's compliance framework 48 - **Deterministic**: Predictable security boundaries, no privilege bypass
49 49
50 ### **Built-In Safety** 50 ### **Security Model**
51 - **Audit Trails**: Every AI action logged and reversible 51 - **Complete Discovery**: See all screens that exist in the system
52 - **Financial Limits**: Can't exceed authorized spending thresholds 52 - **User Enforcement**: Access limited to user's actual permissions
53 - **Multi-Approval**: Critical decisions require human supervisor 53 - **Audit Safe**: All actions logged and traceable to real users
54 - **Container Isolation**: Run in Docker/Kubernetes with separate databases 54 - **Container Isolation**: Run in Docker/Kubernetes with separate databases
55 55
56 **Result**: AI agents follow same career progression and safety protocols as human employees. 56 **Result**: Elegant design - discover everything, enforce user permissions strictly.
57 57
58 ## Overview 58 ## Overview
59 59
......