Ship agents without becoming the governance bottleneck.

Your existing Envoy makes the check. The processor fails closed.

# Filter excerpt; add before envoy.filters.http.router in your existing HTTP connection manager.
- name: envoy.filters.http.ext_proc
  typed_config:
    "@type": type.googleapis.com/envoy.extensions.filters.http.ext_proc.v3.ExternalProcessor
    grpc_service:
      envoy_grpc: { cluster_name: governor_extproc }
    processing_mode:
      request_header_mode: SEND
      request_body_mode: BUFFERED
    failure_mode_allow: false
    allow_mode_override: false
    message_timeout: 5s

Declare a governance object and its namespace assignment in your review workflow.

# Policy object and assignment; provider configured separately.
resource "governor_policy" "production_changes" {
  name        = "production-changes"
  description = "Review consequential production changes"
  enabled     = true
}

resource "governor_policy_assignment" "production" {
  policy_id    = governor_policy.production_changes.id
  namespace_id = governor_namespace.production.id
}

Wrap a function when the code boundary is the right integration point.

# pip install governor-sdk
from governor import governed_fn

@governed_fn("deploy.apply", agent_id="release-agent")
def apply_release(release_id: str) -> dict:
    # Replace this example body with your existing deployment code.
    return {"release_id": release_id, "status": "sample-only"}

# A deny or approval-required decision stops the wrapped function.

Inspect agents and policies through the published GraphQL interface.

# POST /api/v1/graphql, with your tenant authentication.
# Schema: /api/v1/graphql/schema
{
  agents(namespace: "governor.prod", limit: 50) {
    id
    name
    status
  }
  policies(namespace: "governor.prod") {
    id
    name
    version
  }
}