feat(enforcer): enable context-aware policy decisions via custom functions

Policy decisions rely solely on the incoming admission review request body at present. However a more dynamic approach is preferred:
- we should implement custom casbin functions to query the live kubernetes cluster state;
- it will allows policies to verify global constraints such as total pod counts;
- furthermore integrating a kubernetes client within the enforcer helps fetch related resources dynamically;
- caching frequent queries is also recommended to minimize api server load and latency;
- consequently these improvements empower complex policies that require a broader cluster context.

Start by tracing how the enforcer currently evaluates policies from the incoming admission review request. Then identify the existing enforcer entry points and determine how custom Casbin functions, a Kubernetes client, and query caching would fit together. Done means policies can query live cluster state, support global constraints such as total pod counts, and avoid unnecessary API server load.

Written by the indexing model from the issue text.