Document high-level MCP/DI/imperative entry points

Grow the README with the primary MCP builder (WithBedrockGovernance), DI
container (AddBedrockGovernance), and imperative kernel (AddBedrockGuardrailsPolicy
/ AddCloudWatchAudit) entry points that wire up policy, PII sanitization, and
audit together. Add the corresponding changelog message; no new csproj
dependencies (hosting/DI abstractions come transitively via
Microsoft.AgentGovernance and ModelContextProtocol).

Author: GarrettBeatty

.autover/changes/AWS.Bedrock.MAG-initial-preview.json

@@ -7,7 +7,8 @@
         "Initial preview scaffold (0.1.0-preview): package infrastructure and a shared internal guardrail client wrapping ApplyGuardrail. User-facing features (policy backend, audit sink, PII sanitizer, high-level entry points) ship in follow-up preview releases.",
         "Bedrock Guardrails policy backend: an IExternalPolicyBackend that evaluates tool-call context via ApplyGuardrail and denies calls that trip the guardrail. Fails closed on Bedrock/AWS errors by default. Wire it up imperatively with PolicyEngine.AddExternalBackend.",
         "CloudWatch audit sink: subscribes to the toolkit's AuditEmitter and writes governance events to CloudWatch Logs via AWS.Logger.Core, aggregating per-agent/per-policy counters and publishing them as CloudWatch metrics on a configurable flush interval. Delivery is background and non-blocking; sink hiccups can't break the governance loop.",
-        "Bedrock Guardrails PII sanitization for MCP tool output: BedrockGuardrailsSanitizer runs the ANONYMIZE action on tool-result text blocks (30+ PII entity types). GovernedBedrockMcpServerTool decorates an MCP tool so sanitization runs after the toolkit's own scrubbing. Structured (non-text) tool content is passed through unchanged (post-v1 follow-up)."
+        "Bedrock Guardrails PII sanitization for MCP tool output: BedrockGuardrailsSanitizer runs the ANONYMIZE action on tool-result text blocks (30+ PII entity types). GovernedBedrockMcpServerTool decorates an MCP tool so sanitization runs after the toolkit's own scrubbing. Structured (non-text) tool content is passed through unchanged (post-v1 follow-up).",
+        "High-level entry points: WithBedrockGovernance on an MCP server builder, AddBedrockGovernance on IServiceCollection, and GovernanceKernel.AddBedrockGuardrailsPolicy / AddCloudWatchAudit for imperative wiring. BedrockGovernanceOptions carries umbrella Region/Credentials/FailClosed defaults that flow into each feature."
       ]
     }
   ]

src/AWS.Bedrock.MAG/README.md

@@ -1,12 +1,12 @@
 # AWS.Bedrock.MAG
 
-AWS backends for [Microsoft's Agent Governance Toolkit](https://github.com/microsoft/agent-governance-toolkit). This package plugs AWS managed services into the toolkit's extension points so a .NET AI agent can pick up Bedrock-backed policy, PII sanitization, and durable audit with a few lines of configuration.
+AWS backends for [Microsoft's Agent Governance Toolkit](https://github.com/microsoft/agent-governance-toolkit). This package plugs AWS managed services into the toolkit's existing extension points, so a .NET AI agent gets Bedrock-backed policy, PII sanitization, and durable audit with about two lines of configuration.
 
 - **Bedrock Guardrails policy backend**: ML policy evaluation added alongside the toolkit's rule, OPA, and Cedar backends. Fails closed on error.
-- **CloudWatch audit sink**: writes governance events to CloudWatch Logs with aggregated metrics.
 - **Bedrock Guardrails PII sanitization**: redacts or blocks 30+ PII entity types in MCP tool output.
+- **CloudWatch audit sink**: writes governance events to CloudWatch Logs with aggregated metrics.
 
-> Preview (0.1.0). The API may change while the toolkit's extension surface stabilizes. Remaining features (high-level MCP/DI entry points, inline guardrail checks) ship in follow-up preview releases.
+> Preview (0.1.0). The API may change while the toolkit's extension surface stabilizes. Inline guardrail checks (no pre-created Bedrock guardrail required) ship in a follow-up preview release.
 
 ## Install
 
@@ -18,45 +18,46 @@ Targets net8.0. Requires `Microsoft.AgentGovernance` 5.0.0.
 
 ## Use it
 
-### Imperative (you already hold a kernel)
+### MCP server (primary)
 
-Attach the policy backend and audit sink directly to a `GovernanceKernel` you constructed:
+Add `WithBedrockGovernance` after the toolkit's `WithGovernance`:
 
 ```csharp
-var kernel = new GovernanceKernel(new GovernanceOptions { PolicyPaths = { "policies/default.yaml" } });
-
-kernel.PolicyEngine.AddExternalBackend(
-    new BedrockGuardrailsPolicyBackend(new BedrockGuardrailsPolicyOptions
+builder.Services.AddMcpServer()
+    .WithGovernance(o => o.PolicyPaths.Add("policies/default.yaml"))  // Microsoft's toolkit
+    .WithBedrockGovernance(o =>                                        // this package
     {
-        GuardrailId = "gr-abc123",
-    }));
+        o.Policy.GuardrailId    = "gr-abc123";
+        o.EnablePiiSanitization = true;                               // ANONYMIZE tool output
+        o.Audit.LogGroupName    = "/agent-governance/audit";
+    });
+```
+
+### Non-MCP agent (plain DI)
 
-using var audit = new CloudWatchAuditSink(new CloudWatchAuditOptions
+```csharp
+builder.Services.AddBedrockGovernance(o =>
 {
-    LogGroupName    = "/agent-governance/audit",
-    MetricNamespace = "AgentGovernance/Bedrock",
+    o.Policy.GuardrailId    = "gr-abc123";
+    o.Audit.MetricNamespace = "AgentGovernance/Bedrock";
 });
-audit.Subscribe(kernel.AuditEmitter);
+// Requires a GovernanceKernel already in DI (from the toolkit).
 ```
 
-### PII sanitization on MCP tool output
-
-`BedrockGuardrailsSanitizer` runs the ANONYMIZE action on the text blocks of an MCP tool result. The MCP server wiring that plugs it in for you (via `WithBedrockGovernance`) lands in the next preview; in the meantime you can construct the sanitizer directly to feed it tool-result text:
+### Imperative (you already hold a kernel)
 
 ```csharp
-var sanitizer = new BedrockGuardrailsSanitizer(new BedrockSanitizationOptions
-{
-    GuardrailId = "gr-abc123",
-});
+var kernel = new GovernanceKernel(new GovernanceOptions { PolicyPaths = { "policies/default.yaml" } });
 
-var scrubbed = await sanitizer.SanitizeAsync(rawToolText, cancellationToken);
+kernel.AddBedrockGuardrailsPolicy(o => o.GuardrailId = "gr-abc123");
+using var audit = kernel.AddCloudWatchAudit(o => o.LogGroupName = "/agent-governance/audit");
 ```
 
 ## Required IAM
 
 The credentials the agent runs under need:
 
-- `bedrock:ApplyGuardrail` on the guardrail (policy backend and PII sanitization).
+- `bedrock:ApplyGuardrail` on the guardrail (policy and PII sanitization).
 - `logs:CreateLogGroup`, `logs:CreateLogStream`, `logs:PutLogEvents` on the audit log group (audit sink).
 - `cloudwatch:PutMetricData` (audit metrics, when `EmitMetrics` is on).