@superblocksteam/vite-plugin-file-sync 1.99.99-next.2 → 2.0.0-SNAPSHOT.2143e2e95d0107ab

package/dist/ai-service/agent/tools2/tools/start-test-run.js

@@ -0,0 +1,409 @@
+import { z } from "zod";
+import { getLogger } from "../../../../util/logger.js";
+import { sendUserMessageChannel } from "../../../state-machine/helpers/peer.js";
+import { buildTestSubagentSystemPrompt } from "../../subagents/testing/prompt-builder.js";
+import { SubagentType } from "../../subagents/types.js";
+import { processTestStreamChunk } from "../../utils.js";
+import { createToolFactory, ToolCategory, PermissionLevel } from "../types.js";
+const TOOL_EXECUTION_TIMEOUT_MS = 10 * 60 * 1000;
+const EMBED_HOOK_CONFIGS = [
+    {
+        name: "useEmbedEvent",
+        pattern: /\buseEmbedEvent\s*\(\s*["']([^"']+)["']/g,
+        buildInstruction: (values) => `- useEmbedEvent is used for events: ${values.map((e) => `"${e}"`).join(", ")}. Test by calling browser_playwright_action with action: "triggerEmbedEvent" and eventName set to one of these values, then verify the UI reacts correctly.`,
+    },
+    {
+        name: "useEmitEmbedEvent",
+        pattern: /\buseEmitEmbedEvent\s*\(/g,
+        buildInstruction: () => `- useEmitEmbedEvent is used. After an action that should emit an event (e.g., form submit, button click), call browser_playwright_action with action: "getEmbedEvents" and verify events were emitted.`,
+    },
+    {
+        name: "useEmbedProperties",
+        pattern: /\buseEmbedProperties\s*\(/g,
+        buildInstruction: () => `- useEmbedProperties is used. Call browser_playwright_action with action: "setEmbedProperties" and a properties object, then verify the UI reflects the updated values.`,
+    },
+];
+function detectEmbedHooksAndBuildInstruction(fileSystemInterface) {
+    const logger = getLogger();
+    const detected = new Map();
+    try {
+        const currentFiles = fileSystemInterface.getCurrentFiles();
+        for (const fileMeta of Object.values(currentFiles)) {
+            if (!fileMeta.code)
+                continue;
+            for (const config of EMBED_HOOK_CONFIGS) {
+                for (const match of fileMeta.code.matchAll(config.pattern)) {
+                    const values = detected.get(config.name) ?? [];
+                    if (match[1] && !values.includes(match[1])) {
+                        values.push(match[1]);
+                    }
+                    detected.set(config.name, values);
+                }
+            }
+        }
+        if (detected.size > 0) {
+            logger.info(`[TestSubagent] Found embed hooks: ${JSON.stringify(Object.fromEntries(detected))}`);
+        }
+    }
+    catch {
+        logger.warn(`[TestSubagent] Could not check for embed hooks, skipping`);
+        return "";
+    }
+    const instructions = EMBED_HOOK_CONFIGS.filter((config) => detected.has(config.name)).map((config) => config.buildInstruction(detected.get(config.name) ?? []));
+    if (instructions.length === 0) {
+        return "";
+    }
+    return `- Embed hooks: This app uses embed hooks. ALWAYS include one test case for:\n${instructions.join("\n")}`;
+}
+/**
+ * startTestRun - Invokes the Testing Agent subagent to perform end-to-end testing.
+ *
+ * This tool:
+ * 1. Creates a test run with a unique ID
+ * 2. Sends test_run_start message to UI
+ * 3. Invokes the TEST subagent with filtered tools
+ * 4. Suppresses text/reasoning output (all context in tool descriptions)
+ * 5. Streams tool results back to the user
+ */
+export const startTestRunToolFactory = createToolFactory("startTestRun", ({ clark, services, }) => {
+    const logger = getLogger();
+    return {
+        description: `
+Invoke the Testing Agent to perform end-to-end testing of the application.
+
+Use this tool to start a test run when:
+- The plan has enableTesting set to true
+- The user explicitly asks to test (e.g., "test this", "make sure it works", "verify the changes")
+
+The Testing Agent will:
+- Take screenshots to verify UI state
+- Interact with form fields, buttons, etc.
+- Validate that the plan was executed correctly
+- Report test results
+
+**IMPORTANT**: Call this tool when:
+- enableTesting is set on the plan context (REQUIRED)
+- User explicitly asks to test the app
+- After completing a significant feature build
+- When the user says things like "make sure it works", "verify", "test it out"
+
+The Testing Agent has access to Playwright browser automation and can interact with the running application.
+
+**IF TESTS FAIL (status: "failed")**:
+You MUST fix the issues and re-test. Do NOT call \`build_finalize\` until tests pass.
+1. Review the summary to understand what failed
+2. Fix the code to address the failures
+3. Call \`startTestRun\` again to verify your fixes
+4. Repeat until status is "completed" (max 3 attempts)
+`,
+        category: ToolCategory.SUBAGENT,
+        defaultPermission: PermissionLevel.ALLOW,
+        readOnly: true,
+        timeoutMs: TOOL_EXECUTION_TIMEOUT_MS,
+        inputSchema: z.object({
+            header: z
+                .string()
+                .describe("Short, concise header for display (e.g., 'Calculator Test', 'Login Flow', 'Form Validation')"),
+            testDescription: z
+                .string()
+                .describe("Detailed description of what to test for the prompt (e.g., 'Test basic arithmetic operations, special functions, and edge cases') based on the the user prompt and/or the plan to verify"),
+        }),
+        execute: async ({ header, testDescription }, { abortSignal }) => {
+            const testRunId = crypto.randomUUID();
+            const displayHeader = header || "End-to-end testing";
+            // check if planContext.enableTesting is explicitly set to false. if so exit this tool
+            if (clark.context.planContext?.enableTesting === false) {
+                logger.info(`[TestSubagent] Plan context enableTesting is false, exiting tool`);
+                return {
+                    success: true,
+                    testRunId,
+                    status: "completed",
+                    summary: "Testing disabled in plan context",
+                };
+            }
+            logger.info(`[TestSubagent] Test description: ${testDescription}`);
+            // Phase 1: Generate test cases via LLM call (before subagent execution)
+            const testCaseSchema = z.object({
+                testCases: z
+                    .array(z.object({
+                    id: z.string().describe("Unique ID for this test case"),
+                    name: z.string().describe("Short name for the test case"),
+                    description: z
+                        .string()
+                        .describe("Brief description of what this test verifies"),
+                }))
+                    // No array .min()/.max() here: they emit minItems/maxItems, which
+                    // Snowflake Cortex rejects for generateObject (see AGENTS.md). The
+                    // 1-7 count is guided by the describe text and the prompt instead.
+                    .describe("1-7 test cases covering the key functionality to verify. For simpler functionality, generate fewer test cases."),
+            });
+            let generatedTestCases = [];
+            const embedHooksInstruction = detectEmbedHooksAndBuildInstruction(services.fileSystemInterface);
+            try {
+                const { model, providerConfig } = services.llmProvider.modelForTask("classifyPrompt");
+                const testCasesResult = await services.llmClient.generateObject({
+                    model,
+                    schema: testCaseSchema,
+                    context: { mode: "stateless" },
+                    user: {
+                        role: "user",
+                        content: `You are a QA expert. Generate 3-7 test cases to verify the following functionality.
+
+Test Description: ${testDescription}
+
+${clark.context.planContext?.plan ? `Plan Context:\n${clark.context.planContext.plan}` : ""}
+
+Guidelines:
+- Each test case should verify ONE specific piece of functionality
+- Test cases should be feature-level, not click-level (e.g., "Login form submission" not "Click login button")
+- Use unique IDs like "test-1", "test-2", etc.
+- Keep names concise (max 50 chars)
+- Cover both happy path and edge cases where relevant
+- The test cases are meant to be executed in order, so if there are dependencies between test cases, make sure to order the test cases accordingly.
+- Generate 1-3 test cases for simple changes (styling, colors, text)
+- Generate 3-5 test cases for medium complexity (form fields, simple interactions)
+- Generate 5-7 test cases only for complex features (multi-step flows, validation logic)
+- ONLY test what is described - do NOT test unrelated functionality${embedHooksInstruction ? `\n${embedHooksInstruction}` : ""}
+
+Generate test cases that would be useful for verifying this functionality works correctly.`,
+                    },
+                    providerConfig,
+                    retry: { maxAttempts: 2 },
+                    abortSignal,
+                });
+                generatedTestCases = testCasesResult.testCases;
+                logger.info(`[TestSubagent] Generated ${generatedTestCases.length} test cases`);
+            }
+            catch (error) {
+                // If test case generation fails, continue without predefined cases
+                logger.warn(`[TestSubagent] Failed to generate test cases: ${error instanceof Error ? error.message : String(error)}`);
+            }
+            // Prepare test cases with initial status
+            const testCases = generatedTestCases.map((tc) => ({
+                ...tc,
+                status: "not_started",
+            }));
+            // Store test run ID and test cases in context
+            clark.updateContext({
+                currentTestRunId: testRunId,
+                currentTestCases: testCases,
+                testCaseStatuses: new Map(testCases.map((tc) => [tc.id, "not_started"])),
+            });
+            // Send test_run_start message to UI with test cases
+            const sendUserMessage = sendUserMessageChannel(clark, services.chatSessionStore);
+            await sendUserMessage({
+                type: "test_run_start",
+                testRunId,
+                header: displayHeader,
+                testCases: testCases.length > 0 ? testCases : undefined,
+                role: "assistant",
+                content: "",
+            });
+            // Capture interactive elements from the current page for DOM context
+            // This helps the LLM know which elements are available for interaction
+            // without needing to explore the DOM with evaluate()
+            let interactiveElements = [];
+            try {
+                interactiveElements =
+                    await services.mcpServerManager.captureInteractiveElements();
+                logger.info(`[TestSubagent] Captured ${interactiveElements.length} interactive elements for DOM context`);
+            }
+            catch (error) {
+                logger.warn(`[TestSubagent] Failed to capture interactive elements: ${error instanceof Error ? error.message : String(error)}`);
+            }
+            // Build test-specific system prompt with test cases and DOM context
+            const testSystemPrompt = buildTestSubagentSystemPrompt({
+                testRunId,
+                header: displayHeader,
+                planContext: clark.context.planContext,
+                testCases,
+                interactiveElements,
+            });
+            // Get tools filtered for TEST subagent
+            const accessControl = clark.getAccessControl(services.features);
+            const testTools = await services.toolRegistry.getToolsRecord({
+                ...accessControl,
+                subagent: SubagentType.TEST,
+            });
+            // Invoke the test subagent
+            try {
+                const maxSteps = 100;
+                const testCaseList = testCases.length > 0
+                    ? `\n\nYou MUST execute ALL ${testCases.length} test cases listed in the system prompt:\n${testCases.map((tc) => `- ${tc.id}: ${tc.name}`).join("\n")}`
+                    : "";
+                const result = await clark.streamText(services, {
+                    abortSignal,
+                    system: testSystemPrompt,
+                    user: {
+                        role: "user",
+                        content: `Test the application.${testCaseList}\n\nContext: ${testDescription}`,
+                    },
+                    tools: testTools,
+                    context: {
+                        // don't send over the full chat history to improve performance
+                        messages: [],
+                    },
+                    maxSteps,
+                    // Stop when endTestRun is called (like judge stops on submitFeedback)
+                    stopWhen: ({ steps }) => {
+                        const lastStep = steps[steps.length - 1];
+                        const hasEndTestRun = lastStep?.toolCalls?.some((tc) => tc.toolName === "endTestRun");
+                        return hasEndTestRun ?? false;
+                    },
+                    // Warn about remaining steps and force endTestRun on final step
+                    prepareStep: (step) => {
+                        const remainingSteps = maxSteps - step.stepNumber - 1;
+                        const halfwayPoint = Math.floor(maxSteps / 2);
+                        let modifiedStep = step;
+                        // Check if user requested to skip a test case
+                        const skipRequestedForTestCase = clark.context.skipRequestedForTestCase;
+                        if (skipRequestedForTestCase) {
+                            logger.info(`[TestSubagent] User requested skip for test case: ${skipRequestedForTestCase}`);
+                            // Find the next test case to run
+                            const currentTestCases = clark.context.currentTestCases ?? [];
+                            const skippedIndex = currentTestCases.findIndex((tc) => tc.id === skipRequestedForTestCase);
+                            const nextTestCase = skippedIndex >= 0 &&
+                                skippedIndex < currentTestCases.length - 1
+                                ? currentTestCases[skippedIndex + 1]
+                                : null;
+                            let skipInstruction;
+                            if (nextTestCase) {
+                                skipInstruction = `**USER SKIPPED TEST CASE**: The user has skipped test case "${skipRequestedForTestCase}". Move immediately to the next test case "${nextTestCase.id}" (${nextTestCase.name}). Do NOT continue testing the skipped case. Call \`updateTestCaseStatus\` with status "running" for "${nextTestCase.id}" and proceed with testing it.`;
+                            }
+                            else {
+                                skipInstruction = `**USER SKIPPED TEST CASE**: The user has skipped test case "${skipRequestedForTestCase}" which was the last test case. Call \`endTestRun\` to complete the test run.`;
+                            }
+                            const skipMessage = {
+                                role: "user",
+                                content: [
+                                    {
+                                        type: "text",
+                                        text: skipInstruction,
+                                    },
+                                ],
+                            };
+                            // We'll keep injecting this instruction until the agent actually starts a different
+                            // test case (cleared in updateTestCaseStatus) or ends the run.
+                            const lastMessage = modifiedStep.messages.at(-1);
+                            const lastText = typeof lastMessage === "object" &&
+                                lastMessage &&
+                                "content" in lastMessage &&
+                                Array.isArray(lastMessage.content)
+                                ? lastMessage.content?.[0]?.text
+                                : undefined;
+                            modifiedStep =
+                                lastText === skipInstruction
+                                    ? modifiedStep
+                                    : {
+                                        ...modifiedStep,
+                                        messages: [...modifiedStep.messages, skipMessage],
+                                    };
+                        }
+                        // Final step: restrict to endTestRun only
+                        if (step.stepNumber === maxSteps - 1) {
+                            const warningMessage = {
+                                role: "user",
+                                content: [
+                                    {
+                                        type: "text",
+                                        text: "**CRITICAL: This is your FINAL step. You MUST call `endTestRun` now. No other tools are available.**",
+                                    },
+                                ],
+                            };
+                            return {
+                                ...modifiedStep,
+                                messages: [...modifiedStep.messages, warningMessage],
+                                activeTools: ["endTestRun"],
+                            };
+                        }
+                        // Warn from halfway point onwards
+                        else if (step.stepNumber >= halfwayPoint && remainingSteps > 0) {
+                            const warningMessage = {
+                                role: "user",
+                                content: [
+                                    {
+                                        type: "text",
+                                        text: `**IMPORTANT: You have ${remainingSteps} step${remainingSteps === 1 ? "" : "s"} remaining. Wrap up testing and call \`endTestRun\` soon.**`,
+                                    },
+                                ],
+                            };
+                            return {
+                                ...modifiedStep,
+                                messages: [...modifiedStep.messages, warningMessage],
+                            };
+                        }
+                        return modifiedStep;
+                    },
+                });
+                let testSummary = null;
+                // Process stream chunks with test-specific handling
+                // - Text/reasoning is suppressed (not sent to UI)
+                // - browser_playwright_action results become test steps
+                // - endTestRun handles sending test_run_end message with reconciliation
+                for await (const chunk of result.fullStream) {
+                    // Extract summary from endTestRun result for the return value
+                    if (chunk.type === "tool-result" &&
+                        "toolName" in chunk &&
+                        chunk.toolName === "endTestRun") {
+                        const output = chunk.output;
+                        if (output?.success) {
+                            testSummary = {
+                                testStatus: output.testStatus,
+                                summary: output.summary,
+                            };
+                            logger.info(`[TestSubagent] Test summary extracted - ${output.summary}`);
+                        }
+                    }
+                    await processTestStreamChunk(chunk, clark, services);
+                }
+                logger.info(`[TestSubagent] Test run completed: ${testRunId}`);
+                // Build the final result from testSummary, using defaults for missing values
+                const status = testSummary?.testStatus ?? "completed";
+                const summary = testSummary?.summary ?? "Testing completed without summary.";
+                // Note: test_run_end message is sent by endTestRun tool with reconciliation
+                // We need to send it here only if endTestRun wasn't called (shouldn't happen normally)
+                if (!testSummary) {
+                    void sendUserMessage({
+                        type: "test_run_end",
+                        testRunId,
+                        testStatus: "completed",
+                        summary: "Testing completed.",
+                        role: "assistant",
+                        content: "",
+                    });
+                }
+                return {
+                    success: true,
+                    testRunId,
+                    status,
+                    summary,
+                };
+            }
+            catch (error) {
+                logger.error(`[TestSubagent] Test run failed: ${error instanceof Error ? error.message : String(error)}`);
+                // Send test_run_end with failed status
+                await sendUserMessage({
+                    type: "test_run_end",
+                    testRunId,
+                    testStatus: "failed",
+                    role: "assistant",
+                    content: "",
+                });
+                // Clear test run context
+                clark.updateContext({
+                    currentTestRunId: undefined,
+                });
+                throw error;
+            }
+        },
+        getSummary: (input, output) => {
+            const status = output.success ? output.status : "failed";
+            const header = input.header || "test run";
+            return {
+                input: `"${header}"`,
+                output: `Test run ${status}`,
+            };
+        },
+    };
+});
+//# sourceMappingURL=start-test-run.js.map

package/dist/ai-service/agent/tools2/tools/start-test-run.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"start-test-run.js","sourceRoot":"","sources":["../../../../../src/ai-service/agent/tools2/tools/start-test-run.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAIxB,OAAO,EAAE,SAAS,EAAE,MAAM,4BAA4B,CAAC;AAMvD,OAAO,EAAE,sBAAsB,EAAE,MAAM,wCAAwC,CAAC;AAChF,OAAO,EAAE,6BAA6B,EAAE,MAAM,2CAA2C,CAAC;AAC1F,OAAO,EAAE,YAAY,EAAE,MAAM,0BAA0B,CAAC;AACxD,OAAO,EAAE,sBAAsB,EAAE,MAAM,gBAAgB,CAAC;AACxD,OAAO,EAAE,iBAAiB,EAAE,YAAY,EAAE,eAAe,EAAE,MAAM,aAAa,CAAC;AAE/E,MAAM,yBAAyB,GAAG,EAAE,GAAG,EAAE,GAAG,IAAI,CAAC;AAEjD,MAAM,kBAAkB,GAAG;IACzB;QACE,IAAI,EAAE,eAAe;QACrB,OAAO,EAAE,0CAA0C;QACnD,gBAAgB,EAAE,CAAC,MAAgB,EAAE,EAAE,CACrC,uCAAuC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,6JAA6J;KAC7O;IACD;QACE,IAAI,EAAE,mBAAmB;QACzB,OAAO,EAAE,2BAA2B;QACpC,gBAAgB,EAAE,GAAG,EAAE,CACrB,wMAAwM;KAC3M;IACD;QACE,IAAI,EAAE,oBAAoB;QAC1B,OAAO,EAAE,4BAA4B;QACrC,gBAAgB,EAAE,GAAG,EAAE,CACrB,yKAAyK;KAC5K;CACO,CAAC;AAEX,SAAS,mCAAmC,CAC1C,mBAAwC;IAExC,MAAM,MAAM,GAAG,SAAS,EAAE,CAAC;IAC3B,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAoB,CAAC;IAE7C,IAAI,CAAC;QACH,MAAM,YAAY,GAAG,mBAAmB,CAAC,eAAe,EAAE,CAAC;QAE3D,KAAK,MAAM,QAAQ,IAAI,MAAM,CAAC,MAAM,CAAC,YAAY,CAAC,EAAE,CAAC;YACnD,IAAI,CAAC,QAAQ,CAAC,IAAI;gBAAE,SAAS;YAE7B,KAAK,MAAM,MAAM,IAAI,kBAAkB,EAAE,CAAC;gBACxC,KAAK,MAAM,KAAK,IAAI,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,OAAO,CAAC,EAAE,CAAC;oBAC3D,MAAM,MAAM,GAAG,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;oBAC/C,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;wBAC3C,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;oBACxB,CAAC;oBACD,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;gBACpC,CAAC;YACH,CAAC;QACH,CAAC;QAED,IAAI,QAAQ,CAAC,IAAI,GAAG,CAAC,EAAE,CAAC;YACtB,MAAM,CAAC,IAAI,CACT,qCAAqC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,WAAW,CAAC,QAAQ,CAAC,CAAC,EAAE,CACpF,CAAC;QACJ,CAAC;IACH,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,CAAC,IAAI,CAAC,0DAA0D,CAAC,CAAC;QACxE,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,MAAM,YAAY,GAAG,kBAAkB,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CACxD,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAC1B,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,gBAAgB,CAAC,QAAQ,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IAE5E,IAAI,YAAY,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC9B,OAAO,EAAE,CAAC;IACZ,CAAC;IAED,OAAO,gFAAgF,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;AACnH,CAAC;AAED;;;;;;;;;GASG;AACH,MAAM,CAAC,MAAM,uBAAuB,GAAG,iBAAiB,CACtD,cAAc,EACd,CAAC,EACC,KAAK,EACL,QAAQ,GAIT,EAAE,EAAE;IACH,MAAM,MAAM,GAAG,SAAS,EAAE,CAAC;IAE3B,OAAO;QACL,WAAW,EAAE;;;;;;;;;;;;;;;;;;;;;;;;;;;CA2BlB;QACK,QAAQ,EAAE,YAAY,CAAC,QAAQ;QAC/B,iBAAiB,EAAE,eAAe,CAAC,KAAK;QACxC,QAAQ,EAAE,IAAI;QACd,SAAS,EAAE,yBAAyB;QACpC,WAAW,EAAE,CAAC,CAAC,MAAM,CAAC;YACpB,MAAM,EAAE,CAAC;iBACN,MAAM,EAAE;iBACR,QAAQ,CACP,8FAA8F,CAC/F;YACH,eAAe,EAAE,CAAC;iBACf,MAAM,EAAE;iBACR,QAAQ,CACP,0LAA0L,CAC3L;SACJ,CAAC;QACF,OAAO,EAAE,KAAK,EAAE,EAAE,MAAM,EAAE,eAAe,EAAE,EAAE,EAAE,WAAW,EAAE,EAAE,EAAE;YAC9D,MAAM,SAAS,GAAG,MAAM,CAAC,UAAU,EAAE,CAAC;YACtC,MAAM,aAAa,GAAG,MAAM,IAAI,oBAAoB,CAAC;YACrD,sFAAsF;YACtF,IAAI,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,aAAa,KAAK,KAAK,EAAE,CAAC;gBACvD,MAAM,CAAC,IAAI,CACT,kEAAkE,CACnE,CAAC;gBACF,OAAO;oBACL,OAAO,EAAE,IAAI;oBACb,SAAS;oBACT,MAAM,EAAE,WAAW;oBACnB,OAAO,EAAE,kCAAkC;iBAC5C,CAAC;YACJ,CAAC;YAED,MAAM,CAAC,IAAI,CAAC,oCAAoC,eAAe,EAAE,CAAC,CAAC;YAEnE,wEAAwE;YACxE,MAAM,cAAc,GAAG,CAAC,CAAC,MAAM,CAAC;gBAC9B,SAAS,EAAE,CAAC;qBACT,KAAK,CACJ,CAAC,CAAC,MAAM,CAAC;oBACP,EAAE,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,CAAC,8BAA8B,CAAC;oBACvD,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,CAAC,8BAA8B,CAAC;oBACzD,WAAW,EAAE,CAAC;yBACX,MAAM,EAAE;yBACR,QAAQ,CAAC,8CAA8C,CAAC;iBAC5D,CAAC,CACH;oBACD,kEAAkE;oBAClE,mEAAmE;oBACnE,mEAAmE;qBAClE,QAAQ,CACP,gHAAgH,CACjH;aACJ,CAAC,CAAC;YAEH,IAAI,kBAAkB,GAIjB,EAAE,CAAC;YAER,MAAM,qBAAqB,GAAG,mCAAmC,CAC/D,QAAQ,CAAC,mBAAmB,CAC7B,CAAC;YAEF,IAAI,CAAC;gBACH,MAAM,EAAE,KAAK,EAAE,cAAc,EAAE,GAC7B,QAAQ,CAAC,WAAW,CAAC,YAAY,CAAC,gBAAgB,CAAC,CAAC;gBACtD,MAAM,eAAe,GAAG,MAAM,QAAQ,CAAC,SAAS,CAAC,cAAc,CAAC;oBAC9D,KAAK;oBACL,MAAM,EAAE,cAAc;oBACtB,OAAO,EAAE,EAAE,IAAI,EAAE,WAAW,EAAE;oBAC9B,IAAI,EAAE;wBACJ,IAAI,EAAE,MAAM;wBACZ,OAAO,EAAE;;oBAEH,eAAe;;EAEjC,KAAK,CAAC,OAAO,CAAC,WAAW,EAAE,IAAI,CAAC,CAAC,CAAC,kBAAkB,KAAK,CAAC,OAAO,CAAC,WAAW,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE;;;;;;;;;;;;qEAYtB,qBAAqB,CAAC,CAAC,CAAC,KAAK,qBAAqB,EAAE,CAAC,CAAC,CAAC,EAAE;;2FAEnC;qBAC9E;oBACD,cAAc;oBACd,KAAK,EAAE,EAAE,WAAW,EAAE,CAAC,EAAE;oBACzB,WAAW;iBACZ,CAAC,CAAC;gBAEH,kBAAkB,GAAG,eAAe,CAAC,SAAS,CAAC;gBAC/C,MAAM,CAAC,IAAI,CACT,4BAA4B,kBAAkB,CAAC,MAAM,aAAa,CACnE,CAAC;YACJ,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,mEAAmE;gBACnE,MAAM,CAAC,IAAI,CACT,iDAAiD,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAC1G,CAAC;YACJ,CAAC;YAED,yCAAyC;YACzC,MAAM,SAAS,GAAmB,kBAAkB,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC;gBAChE,GAAG,EAAE;gBACL,MAAM,EAAE,aAAsB;aAC/B,CAAC,CAAC,CAAC;YAEJ,8CAA8C;YAC9C,KAAK,CAAC,aAAa,CAAC;gBAClB,gBAAgB,EAAE,SAAS;gBAC3B,gBAAgB,EAAE,SAAS;gBAC3B,gBAAgB,EAAE,IAAI,GAAG,CAGvB,SAAS,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,aAAa,CAAC,CAAC,CAAC;aACjD,CAAC,CAAC;YAEH,oDAAoD;YACpD,MAAM,eAAe,GAAG,sBAAsB,CAC5C,KAAK,EACL,QAAQ,CAAC,gBAAgB,CAC1B,CAAC;YACF,MAAM,eAAe,CAAC;gBACpB,IAAI,EAAE,gBAAgB;gBACtB,SAAS;gBACT,MAAM,EAAE,aAAa;gBACrB,SAAS,EAAE,SAAS,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS;gBACvD,IAAI,EAAE,WAAW;gBACjB,OAAO,EAAE,EAAE;aACZ,CAAC,CAAC;YAEH,qEAAqE;YACrE,uEAAuE;YACvE,qDAAqD;YACrD,IAAI,mBAAmB,GAYlB,EAAE,CAAC;YAER,IAAI,CAAC;gBACH,mBAAmB;oBACjB,MAAM,QAAQ,CAAC,gBAAgB,CAAC,0BAA0B,EAAE,CAAC;gBAC/D,MAAM,CAAC,IAAI,CACT,2BAA2B,mBAAmB,CAAC,MAAM,uCAAuC,CAC7F,CAAC;YACJ,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,MAAM,CAAC,IAAI,CACT,0DAA0D,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CACnH,CAAC;YACJ,CAAC;YAED,oEAAoE;YACpE,MAAM,gBAAgB,GAAG,6BAA6B,CAAC;gBACrD,SAAS;gBACT,MAAM,EAAE,aAAa;gBACrB,WAAW,EAAE,KAAK,CAAC,OAAO,CAAC,WAAW;gBACtC,SAAS;gBACT,mBAAmB;aACpB,CAAC,CAAC;YAEH,uCAAuC;YACvC,MAAM,aAAa,GAAG,KAAK,CAAC,gBAAgB,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;YAEhE,MAAM,SAAS,GAAG,MAAM,QAAQ,CAAC,YAAY,CAAC,cAAc,CAAC;gBAC3D,GAAG,aAAa;gBAChB,QAAQ,EAAE,YAAY,CAAC,IAAI;aAC5B,CAAC,CAAC;YAEH,2BAA2B;YAC3B,IAAI,CAAC;gBACH,MAAM,QAAQ,GAAG,GAAG,CAAC;gBAErB,MAAM,YAAY,GAChB,SAAS,CAAC,MAAM,GAAG,CAAC;oBAClB,CAAC,CAAC,4BAA4B,SAAS,CAAC,MAAM,6CAA6C,SAAS,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,KAAK,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE;oBACvJ,CAAC,CAAC,EAAE,CAAC;gBAET,MAAM,MAAM,GAAG,MAAM,KAAK,CAAC,UAAU,CAAC,QAAQ,EAAE;oBAC9C,WAAW;oBACX,MAAM,EAAE,gBAAgB;oBACxB,IAAI,EAAE;wBACJ,IAAI,EAAE,MAAe;wBACrB,OAAO,EAAE,wBAAwB,YAAY,gBAAgB,eAAe,EAAE;qBAC/E;oBACD,KAAK,EAAE,SAAS;oBAChB,OAAO,EAAE;wBACP,+DAA+D;wBAC/D,QAAQ,EAAE,EAAE;qBACb;oBACD,QAAQ;oBACR,sEAAsE;oBACtE,QAAQ,EAAE,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE;wBACtB,MAAM,QAAQ,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;wBACzC,MAAM,aAAa,GAAG,QAAQ,EAAE,SAAS,EAAE,IAAI,CAC7C,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,QAAQ,KAAK,YAAY,CACrC,CAAC;wBACF,OAAO,aAAa,IAAI,KAAK,CAAC;oBAChC,CAAC;oBACD,gEAAgE;oBAChE,WAAW,EAAE,CAAC,IAAI,EAAE,EAAE;wBACpB,MAAM,cAAc,GAAG,QAAQ,GAAG,IAAI,CAAC,UAAU,GAAG,CAAC,CAAC;wBACtD,MAAM,YAAY,GAAG,IAAI,CAAC,KAAK,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC;wBAC9C,IAAI,YAAY,GAAG,IAAI,CAAC;wBAExB,8CAA8C;wBAC9C,MAAM,wBAAwB,GAC5B,KAAK,CAAC,OAAO,CAAC,wBAAwB,CAAC;wBACzC,IAAI,wBAAwB,EAAE,CAAC;4BAC7B,MAAM,CAAC,IAAI,CACT,qDAAqD,wBAAwB,EAAE,CAChF,CAAC;4BAEF,iCAAiC;4BACjC,MAAM,gBAAgB,GAAG,KAAK,CAAC,OAAO,CAAC,gBAAgB,IAAI,EAAE,CAAC;4BAC9D,MAAM,YAAY,GAAG,gBAAgB,CAAC,SAAS,CAC7C,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,EAAE,KAAK,wBAAwB,CAC3C,CAAC;4BACF,MAAM,YAAY,GAChB,YAAY,IAAI,CAAC;gCACjB,YAAY,GAAG,gBAAgB,CAAC,MAAM,GAAG,CAAC;gCACxC,CAAC,CAAC,gBAAgB,CAAC,YAAY,GAAG,CAAC,CAAC;gCACpC,CAAC,CAAC,IAAI,CAAC;4BAEX,IAAI,eAAuB,CAAC;4BAC5B,IAAI,YAAY,EAAE,CAAC;gCACjB,eAAe,GAAG,+DAA+D,wBAAwB,8CAA8C,YAAY,CAAC,EAAE,MAAM,YAAY,CAAC,IAAI,yGAAyG,YAAY,CAAC,EAAE,gCAAgC,CAAC;4BACxV,CAAC;iCAAM,CAAC;gCACN,eAAe,GAAG,+DAA+D,wBAAwB,+EAA+E,CAAC;4BAC3L,CAAC;4BAED,MAAM,WAAW,GAAG;gCAClB,IAAI,EAAE,MAAe;gCACrB,OAAO,EAAE;oCACP;wCACE,IAAI,EAAE,MAAe;wCACrB,IAAI,EAAE,eAAe;qCACtB;iCACF;6BACF,CAAC;4BAEF,oFAAoF;4BACpF,+DAA+D;4BAC/D,MAAM,WAAW,GAAG,YAAY,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;4BACjD,MAAM,QAAQ,GACZ,OAAO,WAAW,KAAK,QAAQ;gCAC/B,WAAW;gCACX,SAAS,IAAI,WAAW;gCACxB,KAAK,CAAC,OAAO,CAAE,WAAmB,CAAC,OAAO,CAAC;gCACzC,CAAC,CAAE,WAAmB,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,IAAI;gCACzC,CAAC,CAAC,SAAS,CAAC;4BAEhB,YAAY;gCACV,QAAQ,KAAK,eAAe;oCAC1B,CAAC,CAAC,YAAY;oCACd,CAAC,CAAC;wCACE,GAAG,YAAY;wCACf,QAAQ,EAAE,CAAC,GAAG,YAAY,CAAC,QAAQ,EAAE,WAAW,CAAC;qCAClD,CAAC;wBACV,CAAC;wBAED,0CAA0C;wBAC1C,IAAI,IAAI,CAAC,UAAU,KAAK,QAAQ,GAAG,CAAC,EAAE,CAAC;4BACrC,MAAM,cAAc,GAAG;gCACrB,IAAI,EAAE,MAAe;gCACrB,OAAO,EAAE;oCACP;wCACE,IAAI,EAAE,MAAe;wCACrB,IAAI,EAAE,sGAAsG;qCAC7G;iCACF;6BACF,CAAC;4BACF,OAAO;gCACL,GAAG,YAAY;gCACf,QAAQ,EAAE,CAAC,GAAG,YAAY,CAAC,QAAQ,EAAE,cAAc,CAAC;gCACpD,WAAW,EAAE,CAAC,YAAY,CAAC;6BAC5B,CAAC;wBACJ,CAAC;wBACD,kCAAkC;6BAC7B,IAAI,IAAI,CAAC,UAAU,IAAI,YAAY,IAAI,cAAc,GAAG,CAAC,EAAE,CAAC;4BAC/D,MAAM,cAAc,GAAG;gCACrB,IAAI,EAAE,MAAe;gCACrB,OAAO,EAAE;oCACP;wCACE,IAAI,EAAE,MAAe;wCACrB,IAAI,EAAE,yBAAyB,cAAc,QAAQ,cAAc,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,6DAA6D;qCAClJ;iCACF;6BACF,CAAC;4BACF,OAAO;gCACL,GAAG,YAAY;gCACf,QAAQ,EAAE,CAAC,GAAG,YAAY,CAAC,QAAQ,EAAE,cAAc,CAAC;6BACrD,CAAC;wBACJ,CAAC;wBAED,OAAO,YAAY,CAAC;oBACtB,CAAC;iBACF,CAAC,CAAC;gBAEH,IAAI,WAAW,GAGJ,IAAI,CAAC;gBAEhB,oDAAoD;gBACpD,kDAAkD;gBAClD,wDAAwD;gBACxD,wEAAwE;gBACxE,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,CAAC,UAAU,EAAE,CAAC;oBAC5C,8DAA8D;oBAC9D,IACE,KAAK,CAAC,IAAI,KAAK,aAAa;wBAC5B,UAAU,IAAI,KAAK;wBACnB,KAAK,CAAC,QAAQ,KAAK,YAAY,EAC/B,CAAC;wBACD,MAAM,MAAM,GAAI,KAAa,CAAC,MAAM,CAAC;wBACrC,IAAI,MAAM,EAAE,OAAO,EAAE,CAAC;4BACpB,WAAW,GAAG;gCACZ,UAAU,EAAE,MAAM,CAAC,UAAU;gCAC7B,OAAO,EAAE,MAAM,CAAC,OAAO;6BACxB,CAAC;4BACF,MAAM,CAAC,IAAI,CACT,2CAA2C,MAAM,CAAC,OAAO,EAAE,CAC5D,CAAC;wBACJ,CAAC;oBACH,CAAC;oBAED,MAAM,sBAAsB,CAAC,KAAK,EAAE,KAAK,EAAE,QAAQ,CAAC,CAAC;gBACvD,CAAC;gBAED,MAAM,CAAC,IAAI,CAAC,sCAAsC,SAAS,EAAE,CAAC,CAAC;gBAE/D,6EAA6E;gBAC7E,MAAM,MAAM,GAAG,WAAW,EAAE,UAAU,IAAI,WAAW,CAAC;gBACtD,MAAM,OAAO,GACX,WAAW,EAAE,OAAO,IAAI,oCAAoC,CAAC;gBAE/D,4EAA4E;gBAC5E,uFAAuF;gBACvF,IAAI,CAAC,WAAW,EAAE,CAAC;oBACjB,KAAK,eAAe,CAAC;wBACnB,IAAI,EAAE,cAAc;wBACpB,SAAS;wBACT,UAAU,EAAE,WAAW;wBACvB,OAAO,EAAE,oBAAoB;wBAC7B,IAAI,EAAE,WAAW;wBACjB,OAAO,EAAE,EAAE;qBACZ,CAAC,CAAC;gBACL,CAAC;gBAED,OAAO;oBACL,OAAO,EAAE,IAAI;oBACb,SAAS;oBACT,MAAM;oBACN,OAAO;iBACR,CAAC;YACJ,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,MAAM,CAAC,KAAK,CACV,mCAAmC,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAC5F,CAAC;gBAEF,uCAAuC;gBACvC,MAAM,eAAe,CAAC;oBACpB,IAAI,EAAE,cAAc;oBACpB,SAAS;oBACT,UAAU,EAAE,QAAQ;oBACpB,IAAI,EAAE,WAAW;oBACjB,OAAO,EAAE,EAAE;iBACZ,CAAC,CAAC;gBAEH,yBAAyB;gBACzB,KAAK,CAAC,aAAa,CAAC;oBAClB,gBAAgB,EAAE,SAAS;iBAC5B,CAAC,CAAC;gBAEH,MAAM,KAAK,CAAC;YACd,CAAC;QACH,CAAC;QACD,UAAU,EAAE,CAAC,KAAK,EAAE,MAAM,EAAE,EAAE;YAC5B,MAAM,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC;YACzD,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,IAAI,UAAU,CAAC;YAC1C,OAAO;gBACL,KAAK,EAAE,IAAI,MAAM,GAAG;gBACpB,MAAM,EAAE,YAAY,MAAM,EAAE;aAC7B,CAAC;QACJ,CAAC;KACF,CAAC;AACJ,CAAC,CACF,CAAC"}

package/dist/ai-service/agent/tools2/tools/tool-timeout.d.ts

@@ -0,0 +1,2 @@
+export declare const TOOL_EXECUTION_TIMEOUT_BUFFER_MS = 5000;
+//# sourceMappingURL=tool-timeout.d.ts.map

package/dist/ai-service/agent/tools2/tools/tool-timeout.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"tool-timeout.d.ts","sourceRoot":"","sources":["../../../../../src/ai-service/agent/tools2/tools/tool-timeout.ts"],"names":[],"mappings":"AAAA,eAAO,MAAM,gCAAgC,OAAQ,CAAC"}

package/dist/ai-service/agent/tools2/tools/tool-timeout.js

@@ -0,0 +1,2 @@
+export const TOOL_EXECUTION_TIMEOUT_BUFFER_MS = 5_000;
+//# sourceMappingURL=tool-timeout.js.map

package/dist/ai-service/agent/tools2/tools/tool-timeout.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"tool-timeout.js","sourceRoot":"","sources":["../../../../../src/ai-service/agent/tools2/tools/tool-timeout.ts"],"names":[],"mappings":"AAAA,MAAM,CAAC,MAAM,gCAAgC,GAAG,KAAK,CAAC"}

package/dist/ai-service/agent/tools2/tools/update-test-case-status.d.ts

@@ -0,0 +1,29 @@
+import type { Clark, ClarkStateHandlerParams } from "../../../state-machine/clark-fsm.js";
+/**
+ * updateTestCaseStatus - Updates the status of a specific test case during a test run.
+ *
+ * This tool:
+ * 1. Validates that the test case ID exists
+ * 2. Updates the status in clark context (server-side tracking)
+ * 3. Sends test_case_update message to UI
+ *
+ * Only accessible to the TEST subagent (ToolCategory.TEST).
+ */
+export declare const updateTestCaseStatusToolFactory: import("../types.js").ToolFactory<{
+    clark: Clark;
+    services: ClarkStateHandlerParams;
+}, import("../types.js").Tool<{
+    status: "failed" | "running" | "passed" | "skipped";
+    testCaseId: string;
+}, {
+    success: boolean;
+    error: string;
+    testCaseId?: undefined;
+    status?: undefined;
+} | {
+    success: boolean;
+    testCaseId: string;
+    status: "failed" | "running" | "passed" | "skipped";
+    readonly error?: undefined;
+}>>;
+//# sourceMappingURL=update-test-case-status.d.ts.map

package/dist/ai-service/agent/tools2/tools/update-test-case-status.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"update-test-case-status.d.ts","sourceRoot":"","sources":["../../../../../src/ai-service/agent/tools2/tools/update-test-case-status.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,EACV,KAAK,EACL,uBAAuB,EACxB,MAAM,qCAAqC,CAAC;AAI7C;;;;;;;;;GASG;AACH,eAAO,MAAM,+BAA+B;WAMjC,KAAK;cACF,uBAAuB;;;;;;;;;;;;;;GA2HpC,CAAC"}

package/dist/ai-service/agent/tools2/tools/update-test-case-status.js

@@ -0,0 +1,112 @@
+import { z } from "zod";
+import { getLogger } from "../../../../util/logger.js";
+import { sendUserMessageChannel } from "../../../state-machine/helpers/peer.js";
+import { createToolFactory, ToolCategory, PermissionLevel } from "../types.js";
+/**
+ * updateTestCaseStatus - Updates the status of a specific test case during a test run.
+ *
+ * This tool:
+ * 1. Validates that the test case ID exists
+ * 2. Updates the status in clark context (server-side tracking)
+ * 3. Sends test_case_update message to UI
+ *
+ * Only accessible to the TEST subagent (ToolCategory.TEST).
+ */
+export const updateTestCaseStatusToolFactory = createToolFactory("updateTestCaseStatus", ({ clark, services, }) => {
+    const logger = getLogger();
+    return {
+        description: `
+Update the status of a specific test case during the test run.
+
+**IMPORTANT**: You MUST call this tool:
+1. BEFORE starting each test case - set status to "running"
+2. AFTER completing each test case - set status to "passed" or "failed"
+
+Example workflow for test case "test-1":
+1. updateTestCaseStatus({ testCaseId: "test-1", status: "running" })
+2. ... perform browser actions for the test ...
+3. updateTestCaseStatus({ testCaseId: "test-1", status: "passed" })
+
+This ensures the UI shows real-time progress of which tests are running and their results.
+`,
+        category: ToolCategory.TEST,
+        defaultPermission: PermissionLevel.ALLOW,
+        readOnly: true,
+        inputSchema: z.object({
+            testCaseId: z
+                .string()
+                .describe("ID of the test case to update (e.g., 'test-1')"),
+            status: z
+                .enum(["running", "passed", "failed", "skipped"])
+                .describe("New status for the test case: 'running' when starting, 'passed' or 'failed' when complete, 'skipped' when user skips the test"),
+        }),
+        execute: async ({ testCaseId, status }) => {
+            const testRunId = clark.context.currentTestRunId;
+            const testCases = clark.context.currentTestCases;
+            if (!testRunId) {
+                logger.warn(`[TestSubagent] No active test run for updateTestCaseStatus`);
+                return {
+                    success: false,
+                    error: "No active test run",
+                };
+            }
+            const existingStatuses = clark.context.testCaseStatuses;
+            const existingStatus = existingStatuses?.get(testCaseId);
+            if (existingStatus === "skipped" && status !== "skipped") {
+                // Prevent the test subagent from "un-skipping" a test case that the user skipped.
+                // This keeps UI + server-side reconciliation consistent.
+                logger.info(`[TestSubagent] Ignoring status update for skipped test case '${testCaseId}': attempted '${status}'`);
+                return {
+                    success: false,
+                    error: `Test case '${testCaseId}' was skipped by the user`,
+                };
+            }
+            // Validate that the test case exists
+            if (testCases && testCases.length > 0) {
+                const testCaseExists = testCases.some((tc) => tc.id === testCaseId);
+                if (!testCaseExists) {
+                    logger.warn(`[TestSubagent] Test case ID '${testCaseId}' not found in current test cases`);
+                    // Don't fail - allow the update anyway in case of LLM inconsistency
+                }
+            }
+            // Update status in context for server-side tracking
+            const testCaseStatuses = clark.context.testCaseStatuses ??
+                new Map();
+            testCaseStatuses.set(testCaseId, status);
+            // Clear sticky skip instruction once the agent starts a different test case.
+            // This prevents the skip instruction from being injected repeatedly after the agent moves on.
+            const skipRequestedForTestCase = clark.context.skipRequestedForTestCase;
+            clark.updateContext({
+                testCaseStatuses,
+                ...(status === "running" &&
+                    skipRequestedForTestCase &&
+                    testCaseId !== skipRequestedForTestCase
+                    ? { skipRequestedForTestCase: undefined }
+                    : {}),
+            });
+            // Send test_case_update message to UI
+            const sendUserMessage = sendUserMessageChannel(clark, services.chatSessionStore);
+            await sendUserMessage({
+                type: "test_case_update",
+                testRunId,
+                testCaseId,
+                testCaseStatus: status,
+                role: "assistant",
+                content: "",
+            });
+            logger.info(`[TestSubagent] Test case '${testCaseId}' status updated to '${status}'`);
+            return {
+                success: true,
+                testCaseId,
+                status,
+            };
+        },
+        getSummary: (input) => {
+            return {
+                input: `testCaseId: ${input.testCaseId}`,
+                output: `Status: ${input.status}`,
+            };
+        },
+    };
+});
+//# sourceMappingURL=update-test-case-status.js.map

package/dist/ai-service/agent/tools2/tools/update-test-case-status.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"update-test-case-status.js","sourceRoot":"","sources":["../../../../../src/ai-service/agent/tools2/tools/update-test-case-status.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAExB,OAAO,EAAE,SAAS,EAAE,MAAM,4BAA4B,CAAC;AAKvD,OAAO,EAAE,sBAAsB,EAAE,MAAM,wCAAwC,CAAC;AAChF,OAAO,EAAE,iBAAiB,EAAE,YAAY,EAAE,eAAe,EAAE,MAAM,aAAa,CAAC;AAE/E;;;;;;;;;GASG;AACH,MAAM,CAAC,MAAM,+BAA+B,GAAG,iBAAiB,CAC9D,sBAAsB,EACtB,CAAC,EACC,KAAK,EACL,QAAQ,GAIT,EAAE,EAAE;IACH,MAAM,MAAM,GAAG,SAAS,EAAE,CAAC;IAE3B,OAAO;QACL,WAAW,EAAE;;;;;;;;;;;;;CAalB;QACK,QAAQ,EAAE,YAAY,CAAC,IAAI;QAC3B,iBAAiB,EAAE,eAAe,CAAC,KAAK;QACxC,QAAQ,EAAE,IAAI;QACd,WAAW,EAAE,CAAC,CAAC,MAAM,CAAC;YACpB,UAAU,EAAE,CAAC;iBACV,MAAM,EAAE;iBACR,QAAQ,CAAC,gDAAgD,CAAC;YAC7D,MAAM,EAAE,CAAC;iBACN,IAAI,CAAC,CAAC,SAAS,EAAE,QAAQ,EAAE,QAAQ,EAAE,SAAS,CAAC,CAAC;iBAChD,QAAQ,CACP,+HAA+H,CAChI;SACJ,CAAC;QACF,OAAO,EAAE,KAAK,EAAE,EAAE,UAAU,EAAE,MAAM,EAAE,EAAE,EAAE;YACxC,MAAM,SAAS,GAAG,KAAK,CAAC,OAAO,CAAC,gBAAgB,CAAC;YACjD,MAAM,SAAS,GAAG,KAAK,CAAC,OAAO,CAAC,gBAAgB,CAAC;YAEjD,IAAI,CAAC,SAAS,EAAE,CAAC;gBACf,MAAM,CAAC,IAAI,CACT,4DAA4D,CAC7D,CAAC;gBACF,OAAO;oBACL,OAAO,EAAE,KAAK;oBACd,KAAK,EAAE,oBAAoB;iBAC5B,CAAC;YACJ,CAAC;YAED,MAAM,gBAAgB,GAAG,KAAK,CAAC,OAAO,CAAC,gBAAgB,CAAC;YACxD,MAAM,cAAc,GAAG,gBAAgB,EAAE,GAAG,CAAC,UAAU,CAAC,CAAC;YACzD,IAAI,cAAc,KAAK,SAAS,IAAI,MAAM,KAAK,SAAS,EAAE,CAAC;gBACzD,kFAAkF;gBAClF,yDAAyD;gBACzD,MAAM,CAAC,IAAI,CACT,gEAAgE,UAAU,iBAAiB,MAAM,GAAG,CACrG,CAAC;gBACF,OAAO;oBACL,OAAO,EAAE,KAAK;oBACd,KAAK,EAAE,cAAc,UAAU,2BAA2B;iBAC3D,CAAC;YACJ,CAAC;YAED,qCAAqC;YACrC,IAAI,SAAS,IAAI,SAAS,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACtC,MAAM,cAAc,GAAG,SAAS,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,CAAC,EAAE,KAAK,UAAU,CAAC,CAAC;gBACpE,IAAI,CAAC,cAAc,EAAE,CAAC;oBACpB,MAAM,CAAC,IAAI,CACT,gCAAgC,UAAU,mCAAmC,CAC9E,CAAC;oBACF,oEAAoE;gBACtE,CAAC;YACH,CAAC;YAED,oDAAoD;YACpD,MAAM,gBAAgB,GACpB,KAAK,CAAC,OAAO,CAAC,gBAAgB;gBAC9B,IAAI,GAAG,EAGJ,CAAC;YACN,gBAAgB,CAAC,GAAG,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC;YACzC,6EAA6E;YAC7E,8FAA8F;YAC9F,MAAM,wBAAwB,GAAG,KAAK,CAAC,OAAO,CAAC,wBAAwB,CAAC;YACxE,KAAK,CAAC,aAAa,CAAC;gBAClB,gBAAgB;gBAChB,GAAG,CAAC,MAAM,KAAK,SAAS;oBACxB,wBAAwB;oBACxB,UAAU,KAAK,wBAAwB;oBACrC,CAAC,CAAC,EAAE,wBAAwB,EAAE,SAAS,EAAE;oBACzC,CAAC,CAAC,EAAE,CAAC;aACR,CAAC,CAAC;YAEH,sCAAsC;YACtC,MAAM,eAAe,GAAG,sBAAsB,CAC5C,KAAK,EACL,QAAQ,CAAC,gBAAgB,CAC1B,CAAC;YACF,MAAM,eAAe,CAAC;gBACpB,IAAI,EAAE,kBAAkB;gBACxB,SAAS;gBACT,UAAU;gBACV,cAAc,EAAE,MAAM;gBACtB,IAAI,EAAE,WAAW;gBACjB,OAAO,EAAE,EAAE;aACZ,CAAC,CAAC;YAEH,MAAM,CAAC,IAAI,CACT,6BAA6B,UAAU,wBAAwB,MAAM,GAAG,CACzE,CAAC;YAEF,OAAO;gBACL,OAAO,EAAE,IAAI;gBACb,UAAU;gBACV,MAAM;aACP,CAAC;QACJ,CAAC;QACD,UAAU,EAAE,CAAC,KAAK,EAAE,EAAE;YACpB,OAAO;gBACL,KAAK,EAAE,eAAe,KAAK,CAAC,UAAU,EAAE;gBACxC,MAAM,EAAE,WAAW,KAAK,CAAC,MAAM,EAAE;aAClC,CAAC;QACJ,CAAC;KACF,CAAC;AACJ,CAAC,CACF,CAAC"}

package/dist/ai-service/agent/tools2/tools/wait-for-task.d.ts

@@ -0,0 +1,48 @@
+import type { TaskStore } from "../../../tasks/task-store.js";
+/**
+ * Blocks the calling tool turn until the task leaves `running` or the
+ * caller's timeout fires.
+ *
+ * Returns the latest snapshot — including `state: "running"` with
+ * `timedOut: true` — instead of throwing on timeout, so the LLM can
+ * decide whether to keep waiting or do something else.
+ *
+ * Like `checkTask`, the completed-state `result` payload is wired into
+ * `SECRET_SCAN_ENABLED_TOOLS` in `agent/middleware.ts`, so any secrets
+ * a future task type might leak through `result` are redacted before
+ * reaching the LLM. Keep both tools on the allowlist together.
+ */
+export declare const waitForTaskToolFactory: import("../types.js").ToolFactory<{
+    services: {
+        taskStore: TaskStore;
+        getCurrentOwnerId?: () => string | undefined;
+    };
+}, import("../types.js").Tool<{
+    taskId: string;
+    timeoutMs?: number | undefined;
+}, {
+    result: {} | null | undefined;
+    timedOut: false;
+    taskId: string;
+    taskType: string;
+    label?: string;
+    state: import("@superblocksteam/library-shared/types").BackgroundTaskState;
+    message?: string;
+    ageMs: number;
+    completedAt?: string;
+    error?: string;
+    detail?: string;
+} | {
+    result: undefined;
+    timedOut: true;
+    taskId: string;
+    taskType: string;
+    label?: string;
+    state: import("@superblocksteam/library-shared/types").BackgroundTaskState;
+    message?: string;
+    ageMs: number;
+    completedAt?: string;
+    error?: string;
+    detail?: string;
+}>>;
+//# sourceMappingURL=wait-for-task.d.ts.map

package/dist/ai-service/agent/tools2/tools/wait-for-task.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"wait-for-task.d.ts","sourceRoot":"","sources":["../../../../../src/ai-service/agent/tools2/tools/wait-for-task.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,EAAE,SAAS,EAAE,MAAM,8BAA8B,CAAC;AAgB9D;;;;;;;;;;;;GAYG;AACH,eAAO,MAAM,sBAAsB;cAKrB;QACR,SAAS,EAAE,SAAS,CAAC;QACrB,iBAAiB,CAAC,EAAE,MAAM,MAAM,GAAG,SAAS,CAAC;KAC9C;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAqFJ,CAAC"}