Scene 1 — ALLOW: Read-Only Database Query

Policy rule: Scope-based (db.read scope grants access to db.queryReadonly) Expected result: ALLOW

Presentation Rule

Language: All AI narration and communication during scenes must be in English.

The AI must narrate every interaction with the gateway in detail. Before calling the gateway, explain what it is about to do, which endpoint it will call, what payload it will send, and what it expects to happen. After receiving the response, explain what the gateway returned, what the decision means, and why it matters. The goal is to make the audience understand exactly what is happening between the AI and the gateway at every step.

Communication Flow

sequenceDiagram
    participant AI as AI Agent
    participant GW as Gateway
    participant PE as Policy Engine
    participant DB as Database Service

    AI->>GW: POST /api/gateway/execute<br/>{tool: db.queryReadonly, action: query,<br/>arguments: {sql: "SELECT ..."}}
    GW->>GW: Validate JWT + extract scopes
    GW->>PE: Evaluate policy rules
    PE->>PE: Scope check: db.read ✓
    PE-->>GW: Decision: ALLOW (rule: allow-db-readonly)
    GW->>DB: Execute read-only query
    DB-->>GW: Result rows
    GW->>GW: Write audit entry
    GW-->>AI: 200 OK {decision: ALLOW, result: {rows, count}}

Demo Script

Presenter says: "Let's start with a simple, safe operation. The AI wants to read some data from the database."

Action

Example request:

"Query the database to show me all active users"

Send a gateway request for db.queryReadonly with SQL like SELECT * FROM users WHERE active = true.

Expected Response

{
  "request_id": "req-...",
  "decision": "allow",
  "result": {
    "rows": [...],
    "row_count": 5
  }
}

What This Proves

The token has db.read scope -> the policy engine allows read-only queries
The query goes through the gateway, not directly to the database
The AI never sees connection strings or DB credentials
Audit log captures: who queried, what SQL, when, decision=ALLOW

Transition to Next Scene

"OK, reads work fine. But what happens when the AI tries something more dangerous?"

2.2 KiB Raw Permalink Blame History