fix: gateway mode clientId and executePlan endpoint issues

## Issues Fixed

1. **Gateway Mode clientId not sent in request body**
   - `getPolicyApprovedContext()` now auto-populates `client_id` in request body from config
   - Server requires `client_id` in JSON body for `/api/policy/pre-check` endpoint
   - Previously only sent as header (X-Client-ID), causing "client_id field is required" errors
   - Now matches Go SDK behavior

2. **executePlan() using non-existent endpoint**
   - Changed from `/api/v1/orchestrator/plan/{planId}/execute` (404)
   - Now uses `/api/request` with `request_type: "execute-plan"`
   - Matches Go SDK pattern for plan execution
   - Fixes MAP (Multi-Agent Planning) two-step execution

## Testing
- All 80 unit tests pass
- Verified with live examples (PII, SQLi, MAP, Execution Replay)

Author: saurabhjain1592

CHANGELOG.md

@@ -5,6 +5,21 @@ All notable changes to the AxonFlow Java SDK will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [2.1.2] - 2026-01-07
+
+### Fixed
+
+- **Gateway Mode clientId not sent in request body**: Fixed `getPolicyApprovedContext()` to auto-populate `client_id` in request body from config when not explicitly provided
+  - Server requires `client_id` in JSON body for `/api/policy/pre-check` endpoint
+  - Previously only sent as header (X-Client-ID), causing "client_id field is required" errors
+  - Now matches Go SDK behavior which auto-populates from `config.ClientID`
+  - Affects all Gateway Mode pre-check calls
+
+- **executePlan() using non-existent endpoint**: Fixed `executePlan()` to use correct Agent API endpoint
+  - Changed from `/api/v1/orchestrator/plan/{planId}/execute` (404) to `/api/request` with `request_type: "execute-plan"`
+  - Now matches Go SDK pattern for plan execution
+  - Fixes MAP (Multi-Agent Planning) two-step execution flow
+
 ## [2.1.1] - 2026-01-06
 
 ### Fixed

pom.xml

@@ -6,7 +6,7 @@
 
     <groupId>com.getaxonflow</groupId>
     <artifactId>axonflow-sdk</artifactId>
-    <version>2.1.1</version>
+    <version>2.1.2</version>
     <packaging>jar</packaging>
 
     <name>AxonFlow Java SDK</name>

src/main/java/com/getaxonflow/sdk/AxonFlow.java

@@ -261,8 +261,22 @@ public PolicyApprovalResult getPolicyApprovedContext(PolicyApprovalRequest reque
         // Gateway Mode: Let server decide if credentials are required based on DEPLOYMENT_MODE
         // Community/self-hosted deployments work without credentials
 
+        // Auto-populate clientId from config if not set in request (matches Go SDK behavior)
+        PolicyApprovalRequest effectiveRequest = request;
+        if ((request.getClientId() == null || request.getClientId().isEmpty())
+                && config.getClientId() != null && !config.getClientId().isEmpty()) {
+            effectiveRequest = PolicyApprovalRequest.builder()
+                .userToken(request.getUserToken())
+                .query(request.getQuery())
+                .dataSources(request.getDataSources())
+                .context(request.getContext().isEmpty() ? null : request.getContext())
+                .clientId(config.getClientId())
+                .build();
+        }
+
+        final PolicyApprovalRequest finalRequest = effectiveRequest;
         return retryExecutor.execute(() -> {
-            Request httpRequest = buildRequest("POST", "/api/policy/pre-check", request);
+            Request httpRequest = buildRequest("POST", "/api/policy/pre-check", finalRequest);
             try (Response response = httpClient.newCall(httpRequest).execute()) {
                 PolicyApprovalResult result = parseResponse(response, PolicyApprovalResult.class);
 
@@ -648,14 +662,55 @@ public PlanResponse executePlan(String planId) {
         Objects.requireNonNull(planId, "planId cannot be null");
 
         return retryExecutor.execute(() -> {
-            Request httpRequest = buildRequest("POST",
-                "/api/v1/orchestrator/plan/" + planId + "/execute", null);
+            // Build agent request format - like generatePlan but with request_type "execute-plan"
+            String userToken = config.getClientId() != null ? config.getClientId() : "default";
+            String clientId = config.getClientId() != null ? config.getClientId() : "default";
+
+            Map<String, Object> agentRequest = new java.util.HashMap<>();
+            agentRequest.put("query", "");
+            agentRequest.put("user_token", userToken);
+            agentRequest.put("client_id", clientId);
+            agentRequest.put("request_type", "execute-plan");
+            agentRequest.put("context", Map.of("plan_id", planId));
+
+            Request httpRequest = buildRequest("POST", "/api/request", agentRequest);
             try (Response response = httpClient.newCall(httpRequest).execute()) {
-                return parseResponse(response, PlanResponse.class);
+                return parseExecutePlanResponse(response, planId);
             }
         }, "executePlan");
     }
 
+    /**
+     * Parses the execute plan response.
+     */
+    @SuppressWarnings("unchecked")
+    private PlanResponse parseExecutePlanResponse(Response response, String planId) throws IOException {
+        handleErrorResponse(response);
+
+        ResponseBody body = response.body();
+        if (body == null) {
+            throw new AxonFlowException("Empty response body", response.code(), null);
+        }
+
+        String json = body.string();
+        Map<String, Object> agentResponse = objectMapper.readValue(json,
+            new TypeReference<Map<String, Object>>() {});
+
+        // Check for errors
+        Boolean success = (Boolean) agentResponse.get("success");
+        if (success == null || !success) {
+            String error = (String) agentResponse.get("error");
+            throw new PlanExecutionException(error != null ? error : "Plan execution failed");
+        }
+
+        // Extract result - this is the completed plan output
+        String result = (String) agentResponse.get("result");
+
+        // Build response with execution status
+        return new PlanResponse(planId, Collections.emptyList(), null, null, null,
+            null, null, "completed", result);
+    }
+
     /**
      * Gets the status of a plan.
      *

src/test/java/com/getaxonflow/sdk/AxonFlowTest.java

@@ -158,6 +158,69 @@ void getPolicyApprovedContextAsyncShouldReturnFuture() throws Exception {
         assertThat(result.isApproved()).isTrue();
     }
 
+    @Test
+    @DisplayName("getPolicyApprovedContext should auto-populate clientId from config")
+    void getPolicyApprovedContextShouldAutoPopulateClientId(WireMockRuntimeInfo wmRuntimeInfo) {
+        // Create client with clientId configured
+        AxonFlow client = AxonFlow.create(AxonFlowConfig.builder()
+            .endpoint(wmRuntimeInfo.getHttpBaseUrl())
+            .clientId("my-client-id")
+            .clientSecret("my-secret")
+            .build());
+
+        stubFor(post(urlEqualTo("/api/policy/pre-check"))
+            .willReturn(aResponse()
+                .withStatus(200)
+                .withHeader("Content-Type", "application/json")
+                .withBody("{\"context_id\":\"ctx_123\",\"approved\":true}")));
+
+        // Request WITHOUT explicit clientId - SDK should auto-populate from config
+        PolicyApprovalRequest request = PolicyApprovalRequest.builder()
+            .userToken("user-123")
+            .query("What is the capital of France?")
+            .build();
+
+        PolicyApprovalResult result = client.getPolicyApprovedContext(request);
+
+        assertThat(result.isApproved()).isTrue();
+
+        // Verify clientId was sent in request body (server requires this)
+        verify(postRequestedFor(urlEqualTo("/api/policy/pre-check"))
+            .withRequestBody(matchingJsonPath("$.client_id", equalTo("my-client-id"))));
+    }
+
+    @Test
+    @DisplayName("getPolicyApprovedContext should use explicit clientId if provided")
+    void getPolicyApprovedContextShouldUseExplicitClientId(WireMockRuntimeInfo wmRuntimeInfo) {
+        // Create client with clientId configured
+        AxonFlow client = AxonFlow.create(AxonFlowConfig.builder()
+            .endpoint(wmRuntimeInfo.getHttpBaseUrl())
+            .clientId("config-client-id")
+            .clientSecret("my-secret")
+            .build());
+
+        stubFor(post(urlEqualTo("/api/policy/pre-check"))
+            .willReturn(aResponse()
+                .withStatus(200)
+                .withHeader("Content-Type", "application/json")
+                .withBody("{\"context_id\":\"ctx_123\",\"approved\":true}")));
+
+        // Request WITH explicit clientId - should use this one, not config
+        PolicyApprovalRequest request = PolicyApprovalRequest.builder()
+            .userToken("user-123")
+            .query("What is the capital of France?")
+            .clientId("explicit-client-id")
+            .build();
+
+        PolicyApprovalResult result = client.getPolicyApprovedContext(request);
+
+        assertThat(result.isApproved()).isTrue();
+
+        // Verify explicit clientId was sent (not the config one)
+        verify(postRequestedFor(urlEqualTo("/api/policy/pre-check"))
+            .withRequestBody(matchingJsonPath("$.client_id", equalTo("explicit-client-id"))));
+    }
+
     // ========================================================================
     // Gateway Mode - Audit
     // ========================================================================
@@ -257,18 +320,25 @@ void executePlanShouldRequirePlanId() {
     }
 
     @Test
-    @DisplayName("executePlan should execute plan")
+    @DisplayName("executePlan should execute plan via Agent API")
     void executePlanShouldExecutePlan() {
-        stubFor(post(urlEqualTo("/api/v1/orchestrator/plan/plan_123/execute"))
+        // executePlan now uses /api/request with request_type: "execute-plan" (matches Go SDK)
+        stubFor(post(urlEqualTo("/api/request"))
             .willReturn(aResponse()
                 .withStatus(200)
                 .withHeader("Content-Type", "application/json")
-                .withBody("{\"plan_id\":\"plan_123\",\"status\":\"completed\"}")));
+                .withBody("{\"success\":true,\"result\":\"Plan executed successfully\"}")));
 
         PlanResponse response = axonflow.executePlan("plan_123");
 
         assertThat(response.getPlanId()).isEqualTo("plan_123");
         assertThat(response.getStatus()).isEqualTo("completed");
+        assertThat(response.getResult()).isEqualTo("Plan executed successfully");
+
+        // Verify correct request format
+        verify(postRequestedFor(urlEqualTo("/api/request"))
+            .withRequestBody(matchingJsonPath("$.request_type", equalTo("execute-plan")))
+            .withRequestBody(matchingJsonPath("$.context.plan_id", equalTo("plan_123"))));
     }
 
     @Test