Add Microsoft Purview DLP real-time protection on agent-hooks

CHANGELOG.md

@@ -92,6 +92,35 @@ Both `Agent365.Observability.OtelWrite` (Delegated) and `Agent365.Observability.
   agent-hooks interceptor (`ResponsibleAI.AgentHooks` 0.1.0-beta.1) for Defender, and
   `A365AgentHooks.CreateProtectionEmitter` (`parallel/strictest`). When no agent identity is resolved, or resolving
   it fails, the interceptor follows the fail mode.
+- **Microsoft.Agents.A365.Tooling** - Microsoft Purview data loss prevention (DLP) and audit client
+  - `PurviewDlpClient.EvaluateAsync` sends a text to the Microsoft Graph `processContent` API
+    (`POST {base}/me/dataSecurityAndGovernance/processContent`) as `uploadText` (content entering the agent) or
+    `downloadText` (its reply): one conversation entry with the agent, the session (`correlationId`) and its
+    sequence, scoped to the agent's application (the blueprint id by default). A policy action whose
+    `restrictionAction` is `block`, or whose `action` is `blockAccess`, blocks; other actions allow and are counted.
+    Each call sends a new `client-request-id`, which also identifies its content entry and is returned as
+    `PurviewDlpEvaluationResult.CorrelationId`.
+  - Tokens come from a `PurviewDlpTokenResolver`: `PurviewDlpTokenResolvers.FromAgenticUser` uses the agent's
+    connection (`IAgenticTokenProvider.GetAgenticUserTokenAsync`) for the agentic user's delegated Microsoft Graph
+    token (`Content.Process.User`), evaluates as `/me`, and caches the token per tenant, agent, agentic user and scope
+    until it expires (refreshed in the background before it does; failures never cached); `FromAccessTokenProvider`
+    takes a host-supplied token, never cached, for `/me` or for a given user (`/users/{id}`).
+  - Configured with `ENABLE_A365_PURVIEW_DLP`, `A365_PURVIEW_DLP_GRAPH_BASE_URL` (default
+    `https://graph.microsoft.com/v1.0`, absolute HTTPS only), `A365_PURVIEW_DLP_AUTHENTICATION_SCOPE`,
+    `A365_PURVIEW_DLP_FAIL_MODE` (`open` or `closed`), `A365_PURVIEW_DLP_TIMEOUT_MILLISECONDS` (one deadline for token
+    acquisition and the request), `A365_PURVIEW_DLP_MAX_CONTENT_CHARACTERS` (default 100000) and
+    `A365_PURVIEW_DLP_RESPONSE_MODE` (`audit` or `enforce`); any other value is rejected. Failures, and processing
+    errors Graph reports inline, follow the fail mode, recording only the exception type or status, never a response
+    body. Longer text is cut and sent with `isTruncated`: Purview's block stands and its allow follows the fail mode.
+    Redirects are not followed, and lone surrogates are sent as U+FFFD.
+- **Microsoft.Agents.A365.Tooling.Extensions.AgentHooks** - `A365PurviewInterceptor` (`AddA365Purview`), an
+  agent-hooks interceptor for Purview DLP: the user's message at `input` is evaluated as `uploadText` and a block
+  denies it (`purview:block`); the reply at `output` is evaluated as `downloadText`, for audit in the background by
+  default or waiting for the verdict with `A365_PURVIEW_DLP_RESPONSE_MODE=enforce`. Structured content is read only up
+  to `A365_PURVIEW_DLP_MAX_CONTENT_CHARACTERS`. Missing verdicts follow the fail mode (`purview:unverified` or
+  `runtime_error:purview_unverified`). `A365AgentHooks.CreateProtectionEmitter` takes optional `PurviewDlpOptions`,
+  so Defender and Purview compose on one emitter whose interceptor timeout exceeds the longer of their timeouts
+  (Purview's counts only when Purview DLP is enabled).
 - **Microsoft.Agents.A365.Tooling** - V1/V2 per-audience token support for MCP servers
   - `MCPServerConfig` extended with `audience`, `scope`, `publisher`, and `Headers` fields
   - `IMcpTokenProvider` interface for pluggable OAuth token acquisition

docs/design.md

@@ -36,7 +36,7 @@ src/
 │       ├── SemanticKernel/                      # SK tool registration
 │       ├── AgentFramework/                      # AF tool registration
 │       ├── AzureAIFoundry/                      # Azure AI Foundry integration
-│       └── AgentHooks/                          # agent-hooks Defender interceptor (preview)
+│       └── AgentHooks/                          # agent-hooks Defender and Purview interceptors (preview)
 └── Tests/                                       # Test projects
 docs/
 └── design.md                                    # This file
@@ -53,11 +53,11 @@ docs/
 | `Microsoft.Agents.A365.Observability.Extensions.SemanticKernel` | Semantic Kernel telemetry integration | [design.md](../src/Observability/Extensions/SemanticKernel/docs/design.md) |
 | `Microsoft.Agents.A365.Observability.Extensions.OpenAI` | OpenAI SDK telemetry integration | [design.md](../src/Observability/Extensions/OpenAI/docs/design.md) |
 | `Microsoft.Agents.A365.Notifications` | Microsoft 365 notification handling | [design.md](../src/Notification/Microsoft.Agents.A365.Notifications/docs/design.md) |
-| `Microsoft.Agents.A365.Tooling` | MCP server discovery and configuration; Microsoft Defender for AI real-time protection client | [design.md](../src/Tooling/Core/docs/design.md) |
+| `Microsoft.Agents.A365.Tooling` | MCP server discovery and configuration; Microsoft Defender for AI real-time protection and Microsoft Purview DLP clients | [design.md](../src/Tooling/Core/docs/design.md) |
 | `Microsoft.Agents.A365.Tooling.Extensions.SemanticKernel` | Semantic Kernel MCP tool registration | [design.md](../src/Tooling/Extensions/SemanticKernel/docs/design.md) |
 | `Microsoft.Agents.A365.Tooling.Extensions.AgentFramework` | Agent Framework MCP tool registration | [design.md](../src/Tooling/Extensions/AgentFramework/docs/design.md) |
 | `Microsoft.Agents.A365.Tooling.Extensions.AzureAIFoundry` | Azure AI Foundry MCP tool registration | [design.md](../src/Tooling/Extensions/AzureAIFoundry/docs/design.md) |
-| `Microsoft.Agents.A365.Tooling.Extensions.AgentHooks` | Microsoft Defender for AI interceptor for agent-hooks hosts (preview) | [README.md](../src/Tooling/Extensions/AgentHooks/README.md) |
+| `Microsoft.Agents.A365.Tooling.Extensions.AgentHooks` | Microsoft Defender for AI and Microsoft Purview DLP interceptors for agent-hooks hosts (preview) | [README.md](../src/Tooling/Extensions/AgentHooks/README.md) |
 
 ## Core Package Documentation
 
@@ -340,6 +340,10 @@ else if (EnvironmentUtils.IsDevelopmentEnvironment())
 | `ENABLE_A365_DEFENDER_RTP` | Enable Microsoft Defender for AI real-time protection (`DefenderRtpOptions.FromEnvironment`); a value that is not true or false is rejected | `false` |
 | `A365_DEFENDER_RTP_ENDPOINT` | Defender prevention endpoint, `https://<host>/v1/protection/evaluate` | - |
 | `A365_DEFENDER_RTP_FAIL_MODE` | `open` or `closed` (blocks when Defender returns no verdict); any other value is rejected | open |
+| `ENABLE_A365_PURVIEW_DLP` | Enable Microsoft Purview DLP and audit (`PurviewDlpOptions.FromEnvironment`); a value that is not true or false is rejected | `false` |
+| `A365_PURVIEW_DLP_GRAPH_BASE_URL` | Microsoft Graph base URL for `processContent`, an absolute HTTPS URL | `https://graph.microsoft.com/v1.0` |
+| `A365_PURVIEW_DLP_FAIL_MODE` | `open` or `closed` (blocks when Purview returns no verdict); any other value is rejected | open |
+| `A365_PURVIEW_DLP_RESPONSE_MODE` | `audit` (replies are sent to Purview without waiting) or `enforce` (replies wait for the verdict); any other value is rejected | audit |
 
 ## Testing