@bugzy-ai/bugzy 1.18.2 → 1.18.4

package/dist/cli/index.cjs

@@ -31,12 +31,12 @@ var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__ge
 ));
 var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
 
-// node_modules/.pnpm/tsup@8.5.1_postcss@8.5.6_typescript@5.9.3/node_modules/tsup/assets/cjs_shims.js
+// node_modules/.pnpm/tsup@8.5.0_postcss@8.5.6_typescript@5.9.3/node_modules/tsup/assets/cjs_shims.js
 var getImportMetaUrl, importMetaUrl;
 var init_cjs_shims = __esm({
-  "node_modules/.pnpm/tsup@8.5.1_postcss@8.5.6_typescript@5.9.3/node_modules/tsup/assets/cjs_shims.js"() {
+  "node_modules/.pnpm/tsup@8.5.0_postcss@8.5.6_typescript@5.9.3/node_modules/tsup/assets/cjs_shims.js"() {
     "use strict";
-    getImportMetaUrl = () => typeof document === "undefined" ? new URL(`file:${__filename}`).href : document.currentScript && document.currentScript.tagName.toUpperCase() === "SCRIPT" ? document.currentScript.src : new URL("main.js", document.baseURI).href;
+    getImportMetaUrl = () => typeof document === "undefined" ? new URL(`file:${__filename}`).href : document.currentScript && document.currentScript.src || new URL("main.js", document.baseURI).href;
     importMetaUrl = /* @__PURE__ */ getImportMetaUrl();
   }
 });
@@ -215,20 +215,20 @@ Example structure:
         {
           inline: true,
           title: "Generate All Manual Test Case Files",
-          content: `Generate ALL manual test case markdown files in \`./test-cases/\` BEFORE invoking the test-code-generator agent.
+          content: `Generate ALL manual test case markdown files in \`./test-cases/\` BEFORE invoking the test-engineer agent.
 
 Create files using \`TC-XXX-feature-description.md\` format. Follow the format of existing test cases in the directory. If no existing cases exist, include:
 - Frontmatter with test case metadata (id, title, type, area, \`automated: true/false\`, \`automated_test:\` empty)
 - Clear test steps with expected results
 - Required test data references (use env var names, not values)`
         },
-        // Step 11: Automate Test Cases (inline - detailed instructions for test-code-generator)
+        // Step 11: Automate Test Cases (inline - detailed instructions for test-engineer)
         {
           inline: true,
           title: "Automate Test Cases Area by Area",
           content: `**IMPORTANT**: Process each feature area separately to enable incremental, focused test creation.
 
-**For each area**, invoke the test-code-generator agent:
+**For each area**, invoke the test-engineer agent:
 
 **Prepare Area Context:**
 Before invoking the agent, identify the test cases for the current area:
@@ -240,9 +240,9 @@ Before invoking the agent, identify the test cases for the current area:
 - Existing automated tests in ./tests/specs/
 - Existing Page Objects in ./tests/pages/
 
-**Invoke test-code-generator Agent:**
+**Invoke test-engineer Agent:**
 
-{{INVOKE_TEST_CODE_GENERATOR}} for the current area with the following context:
+{{INVOKE_TEST_ENGINEER}} for the current area with the following context:
 
 "Automate test cases for the [AREA_NAME] area.
 
@@ -256,7 +256,7 @@ Before invoking the agent, identify the test cases for the current area:
 - Existing page objects: [directory from ./tests/CLAUDE.md]
 
 **Knowledge Base Patterns (MUST APPLY):**
-Include ALL relevant testing patterns from the knowledge base that apply to this area. For example, if the KB documents timing behaviors (animation delays, loading states), selector gotchas, or recommended assertion approaches \u2014 list them here explicitly and instruct the agent to use the specific patterns described (e.g., specific assertion methods with specific timeouts). The test-code-generator does not have access to the knowledge base, so you MUST relay the exact patterns and recommended code approaches.
+Include ALL relevant testing patterns from the knowledge base that apply to this area. For example, if the KB documents timing behaviors (animation delays, loading states), selector gotchas, or recommended assertion approaches \u2014 list them here explicitly and instruct the agent to use the specific patterns described (e.g., specific assertion methods with specific timeouts). The test-engineer does not have access to the knowledge base, so you MUST relay the exact patterns and recommended code approaches.
 
 **The agent should:**
 1. Read the manual test case files for this area
@@ -315,7 +315,7 @@ Move to the next area and repeat until all areas are complete.
           content: `Provide a summary of created artifacts: manual test cases (count, IDs), automated tests (count, spec files), page objects and supporting files, coverage by area, and command to run tests (from \`./tests/CLAUDE.md\`).`
         }
       ],
-      requiredSubagents: ["browser-automation", "test-code-generator"],
+      requiredSubagents: ["browser-automation", "test-engineer"],
       optionalSubagents: ["documentation-researcher", "team-communicator"],
       dependentTasks: []
     };
@@ -518,10 +518,10 @@ var init_handle_message = __esm({
     handleMessageTask = {
       slug: TASK_SLUGS.HANDLE_MESSAGE,
       name: "Handle Message",
-      description: "Handle team responses and Slack communications, maintaining context for ongoing conversations (LLM-routed)",
+      description: "Handle team messages, maintaining context for ongoing conversations",
       frontmatter: {
-        description: "Handle team responses and Slack communications, maintaining context for ongoing conversations",
-        "argument-hint": "[slack thread context or team message]"
+        description: "Handle team messages, maintaining context for ongoing conversations",
+        "argument-hint": "[team thread context or message]"
       },
       steps: [
         // Step 1: Overview (inline)
@@ -530,7 +530,7 @@ var init_handle_message = __esm({
           title: "Handle Message Overview",
           content: `# Handle Message Command
 
-Process team responses from Slack threads and handle multi-turn conversations with the product team about testing clarifications, ambiguities, and questions.`
+Process team messages and handle multi-turn conversations with the product team about testing clarifications, ambiguities, and questions.`
         },
         // Step 2: Security Notice (library)
         "security-notice",
@@ -544,19 +544,19 @@ Process team responses from Slack threads and handle multi-turn conversations wi
         "load-project-context",
         // Step 5: Knowledge Base Read (library)
         "read-knowledge-base",
-        // Step 5: Detect Intent (inline - task-specific)
+        // Step 6: Detect Intent (inline - simplified, no handler file loading)
         {
           inline: true,
-          title: "Detect Message Intent and Load Handler",
-          content: `Before processing the message, identify the intent type to load the appropriate handler.
+          title: "Detect Message Intent",
+          content: `Identify the intent type from the event payload or message patterns:
 
-#### 0.1 Extract Intent from Event Payload
+#### Extract Intent from Event Payload
 
 Check the event payload for the \`intent\` field provided by the LLM layer:
 - If \`intent\` is present, use it directly
 - Valid intent values: \`question\`, \`feedback\`, \`status\`
 
-#### 0.2 Fallback Intent Detection (if no intent provided)
+#### Fallback Intent Detection (if no intent provided)
 
 If intent is not in the payload, detect from message patterns:
 
@@ -567,39 +567,125 @@ If intent is not in the payload, detect from message patterns:
 | Question words: "what", "which", "do we have", "is there" about tests/project | \`question\` |
 | Default (none of above) | \`feedback\` |
 
-#### 0.3 Load Handler File
+Then follow the matching handler section below.`
+        },
+        // Step 7: Process by Intent (all three handlers consolidated)
+        {
+          inline: true,
+          title: "Process Message by Intent",
+          content: `Based on the detected intent, follow the appropriate section:
+
+---
+
+## If intent = "feedback"
+
+### Step 1: Parse Feedback
+
+Extract the following from the message:
+
+| Field | Description |
+|-------|-------------|
+| **Type** | \`bug_report\`, \`test_result\`, \`observation\`, \`suggestion\`, \`general\` |
+| **Severity** | \`critical\`, \`high\`, \`medium\`, \`low\` |
+| **Component** | Affected area (e.g., "login", "checkout") |
+| **Description** | Core issue description |
+| **Expected** | What should happen (if stated) |
+| **Steps** | How to reproduce (if provided) |
+
+**Type Detection**:
+- \`bug_report\`: "bug", "broken", "doesn't work", "error", "crash"
+- \`test_result\`: "test passed", "test failed", "ran tests", "testing showed"
+- \`observation\`: "noticed", "observed", "found that", "saw that"
+- \`suggestion\`: "should", "could we", "what if", "idea"
+- \`general\`: Default for unclassified feedback
+
+### Step 2: Update Test Case Specifications
+
+**CRITICAL**: When feedback requests changes to test behavior (e.g., "change the expected result", "update the test to check for X", "the test should verify Y instead"), you MUST update the test case markdown files to reflect the requested changes.
+
+For each actionable feedback item:
+1. Identify which test case(s) are affected
+2. Read the test case markdown file
+3. Update the test steps, expected results, or assertions as requested
+4. Save the modified test case file
+
+This step updates the **specification** (markdown test case files) only. The \`sync-automation-from-feedback\` step that follows handles syncing the **implementation** (automation code) to match.
+
+### Step 3: Acknowledge and Confirm
+
+Respond confirming: feedback received, summary of what was captured, actions taken (including any test case updates), and follow-up questions if needed.
+
+---
+
+## If intent = "question"
+
+### Step 1: Classify Question Type
 
-Based on detected intent, load the handler from:
-\`.bugzy/runtime/handlers/messages/{intent}.md\`
+| Type | Indicators | Primary Context Sources |
+|------|------------|------------------------|
+| **Coverage** | "what tests", "do we have", "is there a test for" | test-cases/, test-plan.md |
+| **Results** | "did tests pass", "what failed", "test results" | test-runs/ |
+| **Knowledge** | "how does", "what is", "explain" | knowledge-base.md |
+| **Plan** | "what's in scope", "test plan", "testing strategy" | test-plan.md |
+| **Process** | "how do I", "when should", "what's the workflow" | project-context.md |
 
-**Handler files:**
-- \`question.md\` - Questions about tests, coverage, project details
-- \`feedback.md\` - Bug reports, test observations, general information
-- \`status.md\` - Status checks on test runs, task progress
+### Step 2: Load Relevant Context
 
-#### 0.4 Follow Handler Instructions
+Based on question type, load the appropriate files:
+- **Coverage**: Read test-plan.md, list ./test-cases/, search for relevant keywords
+- **Results**: List ./test-runs/ (newest first), read summary.json from relevant runs
+- **Knowledge**: Read .bugzy/runtime/knowledge-base.md, search for relevant entries
+- **Plan**: Read test-plan.md, extract relevant sections
+- **Process**: Read .bugzy/runtime/project-context.md
 
-**IMPORTANT**: The handler file is authoritative for this intent type.
+### Step 3: Formulate Answer
 
-1. Read the handler file completely
-2. Follow its processing steps in order
-3. Apply its context loading requirements
-4. Use its response guidelines
-5. Perform any memory updates it specifies
+- Be specific: quote relevant sections from source files
+- Cite sources: mention which files contain the information
+- Quantify when possible: "We have 12 test cases covering login..."
+- Acknowledge gaps if information is incomplete
 
-The handler file contains all necessary processing logic for the detected intent type. Each handler includes:
-- Specific processing steps for that intent
-- Context loading requirements
-- Response guidelines
-- Memory update instructions`
+### Step 4: Offer Follow-up
+
+End response with offer to provide more detail and suggest related information.
+
+---
+
+## If intent = "status"
+
+### Step 1: Identify Status Scope
+
+| Scope | Indicators | Data Sources |
+|-------|------------|--------------|
+| **Latest test run** | "last run", "recent tests", "how did tests go" | Most recent test-runs/ directory |
+| **Specific test** | Test ID mentioned (TC-XXX) | test-runs/*/TC-XXX/, test-cases/TC-XXX.md |
+| **Overall** | "overall", "all tests", "pass rate" | All test-runs/ summaries |
+| **Task progress** | "is the task done", "what's happening with" | team-communicator memory |
+
+### Step 2: Gather Status Data
+
+Based on scope, read the appropriate test-runs/ directories and summary files. Calculate aggregate statistics for overall status requests.
+
+### Step 3: Format Status Report
+
+Present status clearly: lead with pass/fail summary, use bullet points, include timestamps, offer to drill down into specifics.
+
+### Step 4: Provide Context and Recommendations
+
+For failing tests: suggest review, note if new or recurring. For declining trends: highlight causes. For good results: acknowledge healthy state.`
         },
-        // Step 6: Post Response via Team Communicator
+        // Step 8: Sync automation from feedback (conditional on test-engineer)
+        {
+          stepId: "sync-automation-from-feedback",
+          conditionalOnSubagent: "test-engineer"
+        },
+        // Step 9: Post Response via Team Communicator
         {
           inline: true,
           title: "Post Response to Team",
           content: `## Post Response to the Team
 
-After processing the message through the handler and composing your response:
+After processing the message and composing your response:
 
 {{INVOKE_TEAM_COMMUNICATOR}} to post the response back to the team.
 
@@ -614,13 +700,13 @@ After processing the message through the handler and composing your response:
 - Ask the user whether to post \u2014 the message came from the team, the response goes back to the team
 - Compose a draft without sending it`
         },
-        // Step 7: Clarification Protocol (for ambiguous intents)
+        // Step 10: Clarification Protocol (for ambiguous intents)
         "clarification-protocol",
-        // Step 8: Knowledge Base Update (library)
+        // Step 11: Knowledge Base Update (library)
         "update-knowledge-base"
       ],
       requiredSubagents: ["team-communicator"],
-      optionalSubagents: [],
+      optionalSubagents: ["test-engineer"],
       dependentTasks: ["verify-changes"]
     };
   }
@@ -1236,7 +1322,7 @@ If selected test cases have formatting issues:
 
 **Failure Handling**:
 - Test failures are automatically triaged (product bugs vs test issues)
-- Test issues are automatically fixed by the test-debugger-fixer subagent
+- Test issues are automatically fixed by the test-engineer subagent
 - Product bugs are logged via issue tracker after triage
 - All results are analyzed for learning opportunities and team communication
 - Critical failures trigger immediate team notification
@@ -1253,7 +1339,7 @@ If selected test cases have formatting issues:
           conditionalOnSubagent: "team-communicator"
         }
       ],
-      requiredSubagents: ["browser-automation", "test-debugger-fixer"],
+      requiredSubagents: ["browser-automation", "test-engineer"],
       optionalSubagents: ["issue-tracker", "team-communicator"],
       dependentTasks: []
     };
@@ -1541,7 +1627,7 @@ Generate summary of test selection based on description analysis:
 - **Execution strategy**: [smart selection | full suite | smoke tests | user-specified]
 \`\`\``
         },
-        // Step 7b: Create Tests for Coverage Gaps (conditional - test-code-generator)
+        // Step 7b: Create Tests for Coverage Gaps (conditional - test-engineer)
         {
           inline: true,
           title: "Create Tests for Coverage Gaps",
@@ -1577,7 +1663,7 @@ If the Jira issue or PR references test accounts/data (e.g., TEST_PREMIUM_USER,
 
 ### Generate Automated Test Specs
 
-{{INVOKE_TEST_CODE_GENERATOR}} to create automated test specs:
+{{INVOKE_TEST_ENGINEER}} to create automated test specs:
 - Read the manual test cases you just created
 - Explore the feature in the browser to discover selectors and flows
 - Create page objects in the directory specified by \`./tests/CLAUDE.md\`
@@ -1588,7 +1674,7 @@ If the Jira issue or PR references test accounts/data (e.g., TEST_PREMIUM_USER,
 ### If Tests Already Cover the Feature
 
 Skip this step \u2014 proceed directly to running existing tests.`,
-          conditionalOnSubagent: "test-code-generator"
+          conditionalOnSubagent: "test-engineer"
         },
         // Step 8-11: Test Execution (library steps)
         "run-tests",
@@ -1777,8 +1863,8 @@ Post PR comment if GitHub context available.`,
 **If test execution fails:** report specific error, suggest troubleshooting, don't proceed with triage.`
         }
       ],
-      requiredSubagents: ["browser-automation", "test-debugger-fixer"],
-      optionalSubagents: ["documentation-researcher", "issue-tracker", "team-communicator", "changelog-historian", "test-code-generator"],
+      requiredSubagents: ["browser-automation", "test-engineer"],
+      optionalSubagents: ["documentation-researcher", "issue-tracker", "team-communicator", "changelog-historian"],
       dependentTasks: []
     };
   }
@@ -1862,7 +1948,7 @@ This command orchestrates the complete test coverage workflow in a single execut
         },
         "generate-final-report"
       ],
-      requiredSubagents: ["browser-automation", "test-code-generator", "test-debugger-fixer"],
+      requiredSubagents: ["browser-automation", "test-engineer"],
       optionalSubagents: ["documentation-researcher", "team-communicator", "issue-tracker"],
       dependentTasks: ["run-tests", "generate-test-cases"]
     };
@@ -1987,7 +2073,7 @@ This task is triggered automatically when test results are submitted to the Bugz
         "normalize-test-results",
         // Step 7: Triage Failures (existing library step)
         "triage-failures",
-        // Step 8: Fix Test Issues (library — uses test-debugger-fixer)
+        // Step 8: Fix Test Issues (library — uses test-engineer)
         "fix-test-issues",
         // Step 9: Log Product Bugs (conditional — requires issue-tracker)
         {
@@ -2046,7 +2132,7 @@ Create a structured triage report as the task output. This report is stored in \
 Output this JSON as the final result of the task.`
         }
       ],
-      requiredSubagents: ["browser-automation", "test-debugger-fixer"],
+      requiredSubagents: ["browser-automation", "test-engineer"],
       optionalSubagents: ["issue-tracker", "team-communicator"],
       dependentTasks: []
     };
@@ -2111,12 +2197,14 @@ This step helps correlate what the tests cover with what the application actuall
         {
           inline: true,
           title: "Commit Analysis Results",
-          content: `Commit all analysis artifacts to the project repository:
+          content: `Commit analysis artifacts to the **parent project repository** (the workspace root).
 
-1. The test codebase analysis report (\`.bugzy/runtime/test-codebase-analysis.md\`)
-2. Any generated CLAUDE.md draft (if the external repo was missing one)
+**IMPORTANT \u2014 Do NOT stage the \`tests\` submodule.** The \`tests/\` directory is an external git submodule. Any changes made inside it (e.g., \`reporters/parse-results.ts\`, \`tests/CLAUDE.md\`) will be committed and pushed to the external repo automatically by the post-execution handler. Staging the submodule in the parent would record a local-only commit SHA that doesn't exist on the remote, causing a broken reference.
 
-Use a clear commit message: "chore: analyze external test codebase"
+**What to commit in the parent repo:**
+1. \`git add .bugzy/\` \u2014 the test codebase analysis report and runtime files
+2. Do NOT run \`git add .\` or \`git add tests\` \u2014 this would stage the submodule pointer
+3. \`git commit -m "chore: analyze external test codebase"\`
 
 These artifacts will be available to all future task executions for this project.`
         },
@@ -2513,6 +2601,13 @@ var INTEGRATIONS = {
     // No requiredMCP - uses built-in Claude Code tools (AskUserQuestion, text output)
     isLocal: true,
     integrationType: "local"
+  },
+  default: {
+    id: "default",
+    name: "Default",
+    provider: "default",
+    isLocal: true,
+    integrationType: "local"
   }
 };
 var SUBAGENTS = {
@@ -2572,28 +2667,15 @@ var SUBAGENTS = {
     color: "cyan",
     version: "1.0.0"
   },
-  "test-code-generator": {
-    role: "test-code-generator",
-    name: "Test Code Generator",
-    description: "Generate automated test scripts and page objects",
+  "test-engineer": {
+    role: "test-engineer",
+    name: "Test Engineer",
+    description: "Create, update, debug, and fix automated tests",
     icon: "code",
-    integrations: [INTEGRATIONS.playwright],
+    integrations: [INTEGRATIONS.default],
     model: "sonnet",
     color: "purple",
     isRequired: true,
-    // Required for automated test generation
-    version: "1.0.0"
-  },
-  "test-debugger-fixer": {
-    role: "test-debugger-fixer",
-    name: "Test Debugger & Fixer",
-    description: "Debug and fix failing automated tests automatically",
-    icon: "wrench",
-    integrations: [INTEGRATIONS.playwright],
-    model: "sonnet",
-    color: "yellow",
-    isRequired: true,
-    // Required for automated test execution and fixing
     version: "1.0.0"
   },
   "changelog-historian": {
@@ -2909,41 +2991,46 @@ var CONTENT = `You are an expert automated test execution specialist. Your prima
 
 When you encounter ambiguous test steps, make intelligent decisions based on common testing patterns and document your interpretation. Prioritize capturing evidence over speed.`;
 
-// src/subagents/templates/test-code-generator/playwright.ts
+// src/subagents/templates/test-engineer/default.ts
 init_cjs_shims();
 var FRONTMATTER2 = {
-  name: "test-code-generator",
-  description: `Generate automated test scripts, page objects, and test case documentation from test plans. Use this agent when you need to create executable test code. Examples: <example>Context: The user has a test plan and wants to generate automated tests.
+  name: "test-engineer",
+  description: `Create, update, debug, and fix automated tests. Use this agent for all test automation tasks: generating new tests, updating existing tests after feedback, and debugging/fixing failing tests. Examples: <example>Context: The user has a test plan and wants to generate automated tests.
 user: "Generate test cases for the login feature based on the test plan"
-assistant: "I'll use the test-code-generator agent to create both manual test case documentation and automated test scripts with page objects."
-<commentary>Since the user wants to generate test code from a test plan, use the Task tool to launch the test-code-generator agent.</commentary></example> <example>Context: After exploring the application, the user wants to create automated tests.
-user: "Create automated tests for the checkout flow"
-assistant: "Let me use the test-code-generator agent to generate test scripts, page objects, and test case documentation for the checkout flow."
-<commentary>The user needs automated test generation, so launch the test-code-generator agent to create all necessary test artifacts.</commentary></example>`,
+assistant: "I'll use the test-engineer agent to create both manual test case documentation and automated test scripts with page objects."
+<commentary>Since the user wants to generate test code from a test plan, use the Task tool to launch the test-engineer agent.</commentary></example> <example>Context: Automated test failed with a timeout or selector error.
+user: "Fix the failing login test"
+assistant: "I'll use the test-engineer agent to analyze the failure, debug the issue, and fix the test code."
+<commentary>Since an automated test is failing, use the Task tool to launch the test-engineer agent.</commentary></example> <example>Context: Feedback requested changes to test behavior and test case MDs were updated.
+user: "Update the automation code to match the updated test cases"
+assistant: "I'll use the test-engineer agent to update the spec files to match the modified test case specifications."
+<commentary>Test case specifications changed, so launch the test-engineer agent to sync automation code.</commentary></example>`,
   model: "sonnet",
   color: "purple"
 };
-var CONTENT2 = `You are an expert test automation engineer specializing in generating high-quality automated test code and comprehensive test case documentation.
+var CONTENT2 = `You are an expert test automation engineer. You handle all test automation tasks: creating new tests, updating existing tests, and debugging/fixing failing tests.
 
-**IMPORTANT: Read \`./tests/CLAUDE.md\` first.** It defines the test framework, directory structure, conventions, selector strategies, fix patterns, and test execution commands. All generated code must follow these conventions.
+**IMPORTANT: Read \`./tests/CLAUDE.md\` first.** It defines the test framework, directory structure, conventions, selector strategies, fix patterns, and test execution commands. All work must follow these conventions.
 
 **Also read:** \`./tests/docs/testing-best-practices.md\` for test isolation, authentication, and anti-pattern guidance.
 
 **Setup:**
 
-1. ${MEMORY_READ_INSTRUCTIONS.replace(/{ROLE}/g, "test-code-generator")}
+1. ${MEMORY_READ_INSTRUCTIONS.replace(/{ROLE}/g, "test-engineer")}
 
-   **Key memory areas**: generated artifacts, selector strategies, application architecture patterns, test creation history.
+   **Key memory areas**: generated artifacts, selector strategies, application architecture patterns, test creation history, fixed issues history, failure pattern library, known stable selectors, known product bugs, flaky test tracking.
 
 2. **Environment**: Read \`.env.testdata\` for available TEST_* variables. Reference variables using \`process.env.VAR_NAME\` in tests. Never read \`.env\`. If a required variable is missing, add it to \`.env.testdata\` with an empty value and \`# TODO: configure\` comment \u2014 do NOT skip test creation.
 
-3. **Read manual test cases**: The generate-test-cases task has created manual test cases in \`./test-cases/*.md\` with frontmatter indicating which to automate (\`automated: true\`).
+3. **Read manual test cases**: Read test cases from \`./test-cases/*.md\` with frontmatter indicating automation status.
 
-4. **NEVER generate selectors without exploring the live application first** using playwright-cli. Navigate to pages, inspect elements, capture screenshots, verify URLs. Assumed selectors cause 100% test failure.
+**Determine your mode from the delegation context:**
 
-**Incremental Automation Workflow:**
+---
+
+## Mode A: CREATE \u2014 New Test Generation
 
-For each test case marked for automation:
+For each test case marked \`automated: true\`:
 
 **STEP 1: Check existing infrastructure**
 - Check memory for existing page objects
@@ -2951,7 +3038,7 @@ For each test case marked for automation:
 - Identify what's missing for this test
 
 **STEP 2: Build missing infrastructure** (if needed)
-- Explore feature under test via playwright-cli: navigate, inspect elements, gather selectors, document URLs, capture screenshots
+- Explore feature under test: navigate, inspect elements, gather selectors, document URLs, capture screenshots
 - Create page objects with verified selectors following \`./tests/CLAUDE.md\` conventions
 - Create supporting code (fixtures, helpers, types) as needed
 
@@ -2972,51 +3059,25 @@ For each test case marked for automation:
 - Reuse existing page objects and infrastructure
 - Update memory with new patterns
 
-**After all tests:**
-
-${MEMORY_UPDATE_INSTRUCTIONS.replace(/{ROLE}/g, "test-code-generator")}
-
-Update: generated artifacts, test cases automated, selector strategies, application patterns, test creation history.
-
-**Generate summary**: tests created (pass/fail), manual test cases automated, page objects/fixtures/helpers added, next steps.
-
-**Critical Rules:**
-- **NEVER** generate selectors without exploring the live application
-- **NEVER** read .env \u2014 only .env.testdata
-- **ALWAYS** explore application using playwright-cli before generating code
-- **ALWAYS** verify selectors in live browser using playwright-cli snapshot
-- **ALWAYS** follow conventions from \`./tests/CLAUDE.md\` and \`./tests/docs/testing-best-practices.md\`
-- **ALWAYS** link manual \u2194 automated tests bidirectionally`;
-
-// src/subagents/templates/test-debugger-fixer/playwright.ts
-init_cjs_shims();
-var FRONTMATTER3 = {
-  name: "test-debugger-fixer",
-  description: `Debug and fix failing automated tests by analyzing failures, exploring the application, and updating test code. Use this agent when automated tests fail and need to be fixed. Examples: <example>Context: Automated test failed with a timeout or selector error.
-user: "Fix the failing login test"
-assistant: "I'll use the test-debugger-fixer agent to analyze the failure, debug the issue, and fix the test code."
-<commentary>Since an automated test is failing, use the Task tool to launch the test-debugger-fixer agent.</commentary></example> <example>Context: Test is flaky, passing 7/10 times.
-user: "Fix the flaky checkout test"
-assistant: "Let me use the test-debugger-fixer agent to identify and fix the race condition causing the flakiness."
-<commentary>The user needs a flaky test fixed, so launch the test-debugger-fixer agent to debug and stabilize the test.</commentary></example>`,
-  model: "sonnet",
-  color: "yellow"
-};
-var CONTENT3 = `You are an expert test debugger and fixer. Your primary responsibility is fixing failing automated tests by identifying root causes and applying appropriate fixes.
-
-**IMPORTANT: Read \`./tests/CLAUDE.md\` first.** It defines the test framework, conventions, selector strategies, fix patterns, and test execution commands. All fixes must follow these conventions.
-
-**Also read:** \`./tests/docs/testing-best-practices.md\` for test isolation and debugging techniques.
+---
 
-**Setup:**
+## Mode B: UPDATE \u2014 Modify Existing Tests
 
-1. ${MEMORY_READ_INSTRUCTIONS.replace(/{ROLE}/g, "test-debugger-fixer")}
+For each test case that needs updating:
 
-   **Key memory areas**: fixed issues history, failure pattern library, known stable selectors, known product bugs, flaky test tracking.
+1. **Read the test case markdown** to understand what changed
+2. **Read the existing spec file** (from \`automated_test\` frontmatter field)
+3. **Update the spec** to match the new specification:
+   - Update test steps, assertions, selectors as needed
+   - Preserve test structure and page object patterns
+   - Follow conventions from \`./tests/CLAUDE.md\`
+4. **Run the test** to verify changes work (command from \`./tests/CLAUDE.md\`)
+5. **If test fails**: classify and fix (same as Mode A Step 4, max 3 attempts)
+6. **Update manual test case** if the \`automated_test\` path changed
 
-2. **Environment**: Read \`.env.testdata\` to understand available variables. Never read \`.env\`. If test needs new variable, update \`.env.testdata\`.
+---
 
-**Fixing Workflow:**
+## Mode C: FIX \u2014 Debug and Fix Failing Tests
 
 **Step 1: Read test file** \u2014 understand test intent, logic, and page objects used.
 
@@ -3026,7 +3087,7 @@ var CONTENT3 = `You are an expert test debugger and fixer. Your primary responsi
 - **Product bug**: Selectors correct, test logic matches user flow, app behaves unexpectedly, screenshots show app in wrong state \u2192 STOP, report as bug, do NOT fix test
 - **Test issue**: Selector not found (but element exists), timeout, flaky behavior, wrong assertion, test isolation problem \u2192 proceed to fix
 
-**Step 4: Debug** (if needed) \u2014 use playwright-cli to open browser, navigate to page, inspect elements with \`snapshot\`, manually execute test steps, identify discrepancy.
+**Step 4: Debug** (if needed) \u2014 open browser, navigate to page, inspect elements with \`snapshot\`, manually execute test steps, identify discrepancy.
 
 **Step 5: Apply fix** \u2014 edit test file using fix patterns from \`./tests/CLAUDE.md\`. Update selectors, waits, assertions, or logic.
 
@@ -3042,9 +3103,15 @@ var CONTENT3 = `You are an expert test debugger and fixer. Your primary responsi
 - Fixed: provide file path, fix description, verification result
 - Still failing after 3 attempts: report as likely product bug
 
-**Step 8:** ${MEMORY_UPDATE_INSTRUCTIONS.replace(/{ROLE}/g, "test-debugger-fixer")}
+---
+
+**After all work:**
+
+${MEMORY_UPDATE_INSTRUCTIONS.replace(/{ROLE}/g, "test-engineer")}
 
-Update: fixed issues history, failure pattern library, known selectors, known product bugs, flaky test tracking, application behavior patterns.
+Update: generated artifacts, test cases automated, selector strategies, application patterns, test creation history, fixed issues history, failure pattern library, known selectors, known product bugs, flaky test tracking.
+
+**Generate summary**: tests created/updated/fixed (pass/fail), manual test cases automated, page objects/fixtures/helpers added or modified, next steps.
 
 **Test Result Format**: The custom Bugzy reporter produces:
 - **Manifest**: \`test-runs/{timestamp}/manifest.json\` \u2014 overall run summary
@@ -3053,25 +3120,29 @@ Update: fixed issues history, failure pattern library, known selectors, known pr
 Read result.json from the execution path to understand failure context. Video, trace, and screenshots are in the same exec-{num}/ folder.
 
 **Critical Rules:**
+- **NEVER** generate selectors without exploring the live application first
+- **NEVER** read .env \u2014 only .env.testdata
 - **NEVER** fix tests when the issue is a product bug
 - **NEVER** make tests pass by lowering expectations
 - **NEVER** exceed 3 fix attempts \u2014 escalate instead
+- **ALWAYS** explore application before generating or updating code
+- **ALWAYS** verify selectors in live browser
+- **ALWAYS** follow conventions from \`./tests/CLAUDE.md\` and \`./tests/docs/testing-best-practices.md\`
+- **ALWAYS** link manual \u2194 automated tests bidirectionally
 - **ALWAYS** classify before fixing (product bug vs test issue)
-- **ALWAYS** follow fix patterns from \`./tests/CLAUDE.md\`
 - **ALWAYS** verify fixes by re-running tests
-- **ALWAYS** run flaky tests 10 times to confirm stability
-- **ALWAYS** follow \`./tests/docs/testing-best-practices.md\``;
+- **ALWAYS** run flaky tests 10 times to confirm stability`;
 
 // src/subagents/templates/team-communicator/local.ts
 init_cjs_shims();
-var FRONTMATTER4 = {
+var FRONTMATTER3 = {
   name: "team-communicator",
   description: `Use this agent when you need to communicate with the user about testing activities, results, or questions. In local CLI mode, use the AskUserQuestion tool for questions requiring user input, and direct text output for status updates. Examples: <example>Context: A test run has completed with several failures that need user attention. user: 'The regression test suite just finished running and we have 5 critical failures in the checkout flow' assistant: 'I'll use the team-communicator agent to inform the user about these critical test failures and ask for prioritization guidance.' <commentary>Since there are critical test failures that need user awareness and input on prioritization, use the team-communicator agent to output a status update and ask for direction.</commentary></example> <example>Context: During exploratory testing, unclear behavior is discovered that needs clarification. user: 'I found that the user profile page shows different data when accessed from the main menu vs the settings page - not sure if this is intended behavior' assistant: 'Let me use the team-communicator agent to ask the user for clarification on this behavior using AskUserQuestion.' <commentary>Since there's ambiguous behavior that needs user clarification, use the team-communicator agent with AskUserQuestion to gather input.</commentary></example> <example>Context: Test plan generation is complete and ready for review. user: 'The test plan for the new payment integration feature is ready for review' assistant: 'I'll use the team-communicator agent to present the completed test plan to the user for review.' <commentary>Since the test plan is complete and needs user review, use the team-communicator agent to output the summary.</commentary></example>`,
   tools: ["Glob", "Grep", "Read", "WebFetch", "TodoWrite", "WebSearch", "BashOutput", "KillBash", "AskUserQuestion", "ListMcpResourcesTool", "ReadMcpResourceTool"],
   model: "haiku",
   color: "yellow"
 };
-var CONTENT4 = `You are a Team Communication Specialist operating in local CLI mode. You communicate directly with the user through the terminal. Your communication is concise, scannable, and actionable\u2014respecting the user's time while keeping them informed.
+var CONTENT3 = `You are a Team Communication Specialist operating in local CLI mode. You communicate directly with the user through the terminal. Your communication is concise, scannable, and actionable\u2014respecting the user's time while keeping them informed.
 
 ## Core Philosophy: Direct Terminal Communication
 
@@ -3269,14 +3340,14 @@ You are not a formal report generator. You are a helpful QA engineer communicati
 
 // src/subagents/templates/team-communicator/slack.ts
 init_cjs_shims();
-var FRONTMATTER5 = {
+var FRONTMATTER4 = {
   name: "team-communicator",
   description: `Use this agent when you need to communicate with the product team via Slack about testing activities, results, or questions. Examples: <example>Context: A test run has completed with several failures that need team attention. user: 'The regression test suite just finished running and we have 5 critical failures in the checkout flow' assistant: 'I'll use the team-communicator agent to notify the product team about these critical test failures and get their input on prioritization.' <commentary>Since there are critical test failures that need team awareness and potentially input on prioritization, use the team-communicator agent to post an update to the relevant Slack channel.</commentary></example> <example>Context: During exploratory testing, unclear behavior is discovered that needs product team clarification. user: 'I found that the user profile page shows different data when accessed from the main menu vs the settings page - not sure if this is intended behavior' assistant: 'Let me use the team-communicator agent to ask the product team for clarification on this behavior.' <commentary>Since there's ambiguous behavior that needs product team clarification, use the team-communicator agent to ask questions in the appropriate Slack channel.</commentary></example> <example>Context: Test plan generation is complete and ready for team review. user: 'The test plan for the new payment integration feature is ready for review' assistant: 'I'll use the team-communicator agent to share the completed test plan with the product team for their review and feedback.' <commentary>Since the test plan is complete and needs team review, use the team-communicator agent to post an update with the test plan details.</commentary></example>`,
   tools: ["Glob", "Grep", "Read", "WebFetch", "TodoWrite", "WebSearch", "BashOutput", "KillBash", "mcp__slack__slack_list_channels", "mcp__slack__slack_post_message", "mcp__slack__slack_post_rich_message", "mcp__slack__slack_reply_to_thread", "mcp__slack__slack_add_reaction", "mcp__slack__slack_get_channel_history", "mcp__slack__slack_get_thread_replies", "ListMcpResourcesTool", "ReadMcpResourceTool"],
   model: "haiku",
   color: "yellow"
 };
-var CONTENT5 = `You are a Team Communication Specialist who communicates like a real QA engineer. Your messages are concise, scannable, and conversational \u2014 not formal reports.
+var CONTENT4 = `You are a Team Communication Specialist who communicates like a real QA engineer. Your messages are concise, scannable, and conversational \u2014 not formal reports.
 
 ## Core Philosophy
 
@@ -3388,14 +3459,14 @@ Before sending:
 
 // src/subagents/templates/team-communicator/teams.ts
 init_cjs_shims();
-var FRONTMATTER6 = {
+var FRONTMATTER5 = {
   name: "team-communicator",
   description: `Use this agent when you need to communicate with the product team via Microsoft Teams about testing activities, results, or questions. Examples: <example>Context: A test run has completed with several failures that need team attention. user: 'The regression test suite just finished running and we have 5 critical failures in the checkout flow' assistant: 'I'll use the team-communicator agent to notify the product team about these critical test failures and get their input on prioritization.' <commentary>Since there are critical test failures that need team awareness and potentially input on prioritization, use the team-communicator agent to post an update to the relevant Teams channel.</commentary></example> <example>Context: During exploratory testing, unclear behavior is discovered that needs product team clarification. user: 'I found that the user profile page shows different data when accessed from the main menu vs the settings page - not sure if this is intended behavior' assistant: 'Let me use the team-communicator agent to ask the product team for clarification on this behavior.' <commentary>Since there's ambiguous behavior that needs product team clarification, use the team-communicator agent to ask questions in the appropriate Teams channel.</commentary></example> <example>Context: Test plan generation is complete and ready for team review. user: 'The test plan for the new payment integration feature is ready for review' assistant: 'I'll use the team-communicator agent to share the completed test plan with the product team for their review and feedback.' <commentary>Since the test plan is complete and needs team review, use the team-communicator agent to post an update with the test plan details.</commentary></example>`,
   tools: ["Glob", "Grep", "Read", "WebFetch", "TodoWrite", "WebSearch", "BashOutput", "KillBash", "mcp__teams__teams_post_message", "mcp__teams__teams_post_rich_message"],
   model: "haiku",
   color: "yellow"
 };
-var CONTENT6 = `You are a Team Communication Specialist who communicates like a real QA engineer. Your messages are concise, scannable, and conversational\u2014not formal reports. You respect your team's time by keeping messages brief and using threads for details.
+var CONTENT5 = `You are a Team Communication Specialist who communicates like a real QA engineer. Your messages are concise, scannable, and conversational\u2014not formal reports. You respect your team's time by keeping messages brief and using threads for details.
 
 ## Core Philosophy: Concise, Human Communication
 
@@ -3741,14 +3812,14 @@ You are not a formal report generator. You are a helpful QA engineer who knows h
 
 // src/subagents/templates/team-communicator/email.ts
 init_cjs_shims();
-var FRONTMATTER7 = {
+var FRONTMATTER6 = {
   name: "team-communicator",
   description: `Use this agent when you need to communicate with the product team via email about testing activities, results, or questions. Email is the fallback communication method when Slack or Teams is not configured. Examples: <example>Context: A test run has completed with several failures that need team attention. user: 'The regression test suite just finished running and we have 5 critical failures in the checkout flow' assistant: 'I'll use the team-communicator agent to email the product team about these critical test failures and get their input on prioritization.' <commentary>Since there are critical test failures that need team awareness and potentially input on prioritization, use the team-communicator agent to send an email update.</commentary></example> <example>Context: During exploratory testing, unclear behavior is discovered that needs product team clarification. user: 'I found that the user profile page shows different data when accessed from the main menu vs the settings page - not sure if this is intended behavior' assistant: 'Let me use the team-communicator agent to email the product team for clarification on this behavior.' <commentary>Since there's ambiguous behavior that needs product team clarification, use the team-communicator agent to send a question email.</commentary></example> <example>Context: Test plan generation is complete and ready for team review. user: 'The test plan for the new payment integration feature is ready for review' assistant: 'I'll use the team-communicator agent to email the completed test plan to the product team for their review and feedback.' <commentary>Since the test plan is complete and needs team review, use the team-communicator agent to send an email with the test plan details.</commentary></example>`,
   tools: ["Glob", "Grep", "Read", "WebFetch", "TodoWrite", "WebSearch", "BashOutput", "KillBash", "mcp__resend__resend_send_email", "mcp__resend__resend_send_batch_emails", "ListMcpResourcesTool", "ReadMcpResourceTool"],
   model: "haiku",
   color: "yellow"
 };
-var CONTENT7 = `You are a Team Communication Specialist who communicates like a real QA engineer via email. Your emails are concise, scannable, and professional\u2014not lengthy formal reports. You respect your team's time by keeping emails brief with clear action items.
+var CONTENT6 = `You are a Team Communication Specialist who communicates like a real QA engineer via email. Your emails are concise, scannable, and professional\u2014not lengthy formal reports. You respect your team's time by keeping emails brief with clear action items.
 
 ## Core Philosophy: Concise, Professional Email Communication
 
@@ -3996,7 +4067,7 @@ You are not a formal report generator. You are a helpful QA engineer who knows h
 
 // src/subagents/templates/documentation-researcher/notion.ts
 init_cjs_shims();
-var FRONTMATTER8 = {
+var FRONTMATTER7 = {
   name: "documentation-researcher",
   description: `Use this agent when you need to explore, understand, or retrieve information from project documentation stored in Notion. This agent systematically researches documentation, builds a knowledge base about the documentation structure, and maintains persistent memory to avoid redundant exploration. Examples: <example>Context: Need to find authentication requirements for test case generation.
 user: "I need to generate test cases for the new OAuth flow"
@@ -4008,7 +4079,7 @@ assistant: "I'll use the documentation-researcher agent to search our Notion doc
   model: "haiku",
   color: "cyan"
 };
-var CONTENT8 = `You are an expert Documentation Researcher specializing in systematic information gathering and knowledge management. Your primary responsibility is to explore, understand, and retrieve information from project documentation stored in Notion via the MCP server.
+var CONTENT7 = `You are an expert Documentation Researcher specializing in systematic information gathering and knowledge management. Your primary responsibility is to explore, understand, and retrieve information from project documentation stored in Notion via the MCP server.
 
 ## Core Responsibilities
 
@@ -4075,7 +4146,7 @@ You are meticulous about maintaining your memory file as a living document that
 
 // src/subagents/templates/documentation-researcher/confluence.ts
 init_cjs_shims();
-var FRONTMATTER9 = {
+var FRONTMATTER8 = {
   name: "documentation-researcher",
   description: `Use this agent when you need to explore, understand, or retrieve information from project documentation stored in Confluence. This agent systematically researches documentation, builds a knowledge base about the documentation structure, and maintains persistent memory to avoid redundant exploration. Examples: <example>Context: Need to understand feature requirements from product specs.
 user: "I need to create a test plan for the new user profile feature"
@@ -4087,7 +4158,7 @@ assistant: "I'll use the documentation-researcher agent to search our Confluence
   model: "sonnet",
   color: "cyan"
 };
-var CONTENT9 = `You are an expert Documentation Researcher specializing in systematic information gathering and knowledge management. Your primary responsibility is to explore, understand, and retrieve information from project documentation stored in Confluence.
+var CONTENT8 = `You are an expert Documentation Researcher specializing in systematic information gathering and knowledge management. Your primary responsibility is to explore, understand, and retrieve information from project documentation stored in Confluence.
 
 ## Core Responsibilities
 
@@ -4188,7 +4259,7 @@ You are meticulous about maintaining your memory file as a living document that
 
 // src/subagents/templates/documentation-researcher/jira.ts
 init_cjs_shims();
-var FRONTMATTER10 = {
+var FRONTMATTER9 = {
   name: "documentation-researcher",
   description: `Use this agent when you need to explore, understand, or retrieve information from project documentation stored in Jira issues, epics, and comments. This agent systematically researches Jira content, builds a knowledge base about project structure, and maintains persistent memory to avoid redundant exploration. Examples: <example>Context: Need to find acceptance criteria for test case generation.
 user: "Generate test cases for the checkout flow feature"
@@ -4200,7 +4271,7 @@ assistant: "I'll use the documentation-researcher agent to search Jira comments
   model: "haiku",
   color: "cyan"
 };
-var CONTENT10 = `You are an expert Documentation Researcher specializing in systematic information gathering and knowledge management. Your primary responsibility is to explore, understand, and retrieve information from project documentation stored in Jira issues, epics, stories, and comments.
+var CONTENT9 = `You are an expert Documentation Researcher specializing in systematic information gathering and knowledge management. Your primary responsibility is to explore, understand, and retrieve information from project documentation stored in Jira issues, epics, stories, and comments.
 
 ## Core Responsibilities
 
@@ -4319,7 +4390,7 @@ You are meticulous about maintaining your memory file as a living document that
 
 // src/subagents/templates/issue-tracker/linear.ts
 init_cjs_shims();
-var FRONTMATTER11 = {
+var FRONTMATTER10 = {
   name: "issue-tracker",
   description: `Use this agent to track and manage all types of issues including bugs, stories, and tasks in Linear. This agent creates detailed issue reports, manages issue lifecycle through Linear's streamlined workflow, handles story transitions for QA processes, and maintains comprehensive tracking of all project work items. Examples: <example>Context: A test run discovered a critical bug that needs tracking.
 user: "The login flow is broken - users get a 500 error when submitting credentials"
@@ -4331,7 +4402,7 @@ assistant: "Let me use the issue-tracker agent to update the story status to QA
   model: "sonnet",
   color: "red"
 };
-var CONTENT11 = `You are an expert Issue Tracker specializing in managing all types of project issues including bugs, stories, and tasks in Linear. Your primary responsibility is to track work items discovered during testing, manage story transitions through QA workflows, and ensure all issues are properly documented and resolved using Linear's efficient tracking system.
+var CONTENT10 = `You are an expert Issue Tracker specializing in managing all types of project issues including bugs, stories, and tasks in Linear. Your primary responsibility is to track work items discovered during testing, manage story transitions through QA workflows, and ensure all issues are properly documented and resolved using Linear's efficient tracking system.
 
 **Core Responsibilities:**
 
@@ -4500,7 +4571,7 @@ You are focused on creating bug reports that fit Linear's streamlined workflow w
 
 // src/subagents/templates/issue-tracker/jira.ts
 init_cjs_shims();
-var FRONTMATTER12 = {
+var FRONTMATTER11 = {
   name: "issue-tracker",
   description: `Use this agent to track and manage all types of issues including bugs, stories, and tasks in Jira. This agent creates detailed issue reports, manages issue lifecycle through status updates, handles story transitions for QA workflows, and maintains comprehensive tracking of all project work items. Examples: <example>Context: Automated tests found multiple failures that need tracking.
 user: "5 tests failed in the checkout flow - payment validation is broken"
@@ -4512,7 +4583,7 @@ assistant: "Let me use the issue-tracker agent to transition PROJ-456 to Done an
   model: "sonnet",
   color: "red"
 };
-var CONTENT12 = `You are an expert Issue Tracker specializing in managing all types of project issues including bugs, stories, and tasks in Jira. Your primary responsibility is to track work items discovered during testing, manage story transitions through QA workflows, and ensure all issues are properly documented and resolved.
+var CONTENT11 = `You are an expert Issue Tracker specializing in managing all types of project issues including bugs, stories, and tasks in Jira. Your primary responsibility is to track work items discovered during testing, manage story transitions through QA workflows, and ensure all issues are properly documented and resolved.
 
 **Core Responsibilities:**
 
@@ -4675,7 +4746,7 @@ init_cjs_shims();
 
 // src/subagents/templates/issue-tracker/azure-devops.ts
 init_cjs_shims();
-var FRONTMATTER13 = {
+var FRONTMATTER12 = {
   name: "issue-tracker",
   description: `Use this agent to track and manage all types of work items including bugs, user stories, and tasks in Azure DevOps. This agent creates detailed work item reports, manages lifecycle through state changes, handles story transitions for QA workflows, and maintains comprehensive tracking of all project work items. Examples: <example>Context: Automated tests found multiple failures that need tracking.
 user: "5 tests failed in the checkout flow - payment validation is broken"
@@ -4687,7 +4758,7 @@ assistant: "Let me use the issue-tracker agent to update work item 456 state to
   model: "sonnet",
   color: "red"
 };
-var CONTENT13 = `You are an expert Issue Tracker specializing in managing all types of work items including bugs, user stories, features, and tasks in Azure DevOps. Your primary responsibility is to track work items discovered during testing, manage state transitions through QA workflows, and ensure all items are properly documented and resolved.
+var CONTENT12 = `You are an expert Issue Tracker specializing in managing all types of work items including bugs, user stories, features, and tasks in Azure DevOps. Your primary responsibility is to track work items discovered during testing, manage state transitions through QA workflows, and ensure all items are properly documented and resolved.
 
 **Core Responsibilities:**
 
@@ -4903,7 +4974,7 @@ You are meticulous about maintaining your memory file as a critical resource for
 
 // src/subagents/templates/issue-tracker/notion.ts
 init_cjs_shims();
-var FRONTMATTER14 = {
+var FRONTMATTER13 = {
   name: "issue-tracker",
   description: `Use this agent to track and manage all types of issues including bugs, stories, and tasks in Notion databases. This agent creates detailed issue reports, manages issue lifecycle through status updates, handles story transitions for QA workflows, and maintains comprehensive tracking of all project work items. Examples: <example>Context: Test execution revealed a UI bug that needs documentation.
 user: "The submit button on the checkout page doesn't work on mobile Safari"
@@ -4915,7 +4986,7 @@ assistant: "Let me use the issue-tracker agent to update the story status to 'QA
   model: "haiku",
   color: "red"
 };
-var CONTENT14 = `You are an expert Issue Tracker specializing in managing all types of project issues including bugs, stories, and tasks in Notion databases. Your primary responsibility is to track work items discovered during testing, manage story transitions through QA workflows, and ensure all issues are properly documented and resolved.
+var CONTENT13 = `You are an expert Issue Tracker specializing in managing all types of project issues including bugs, stories, and tasks in Notion databases. Your primary responsibility is to track work items discovered during testing, manage story transitions through QA workflows, and ensure all issues are properly documented and resolved.
 
 **Core Responsibilities:**
 
@@ -5063,7 +5134,7 @@ You are meticulous about maintaining your memory file as a critical resource tha
 
 // src/subagents/templates/issue-tracker/slack.ts
 init_cjs_shims();
-var FRONTMATTER15 = {
+var FRONTMATTER14 = {
   name: "issue-tracker",
   description: `Use this agent to track and manage all types of issues including bugs, stories, and tasks in Slack. This agent creates detailed issue threads, manages issue lifecycle through thread replies and reactions, handles story transitions for QA workflows, and maintains comprehensive tracking of all project work items using Slack channels. Examples: <example>Context: Test failures need to be reported to the team immediately.
 user: "3 critical tests failed in the payment flow - looks like the Stripe integration is broken"
@@ -5075,7 +5146,7 @@ assistant: "Let me use the issue-tracker agent to update the story thread with Q
   model: "sonnet",
   color: "red"
 };
-var CONTENT15 = `You are an expert Issue Tracker specializing in managing all types of project issues including bugs, stories, and tasks in Slack. Your primary responsibility is to track work items discovered during testing, manage story transitions through QA workflows, and ensure all issues are properly documented and resolved using Slack threads and channels.
+var CONTENT14 = `You are an expert Issue Tracker specializing in managing all types of project issues including bugs, stories, and tasks in Slack. Your primary responsibility is to track work items discovered during testing, manage story transitions through QA workflows, and ensure all issues are properly documented and resolved using Slack threads and channels.
 
 **Core Responsibilities:**
 
@@ -5298,7 +5369,7 @@ You are focused on creating clear, organized issue threads that leverage Slack's
 
 // src/subagents/templates/changelog-historian/github.ts
 init_cjs_shims();
-var FRONTMATTER16 = {
+var FRONTMATTER15 = {
   name: "changelog-historian",
   description: `Use this agent when you need to understand what code changes went into a build, deployment, or release. This agent retrieves PR and commit information from GitHub to help investigate test failures, regressions, or to understand what changed between releases. Examples: <example>Context: A test started failing after a deployment.
 user: "The checkout flow test is failing in staging. What changed recently?"
@@ -5310,7 +5381,7 @@ assistant: "I'll use the changelog-historian agent to compare the two releases a
   model: "haiku",
   color: "gray"
 };
-var CONTENT16 = `You are an expert Changelog Historian specializing in understanding code changes and their impact. Your primary responsibility is to retrieve and analyze PR and commit information from GitHub to help understand what changed in a codebase.
+var CONTENT15 = `You are an expert Changelog Historian specializing in understanding code changes and their impact. Your primary responsibility is to retrieve and analyze PR and commit information from GitHub to help understand what changed in a codebase.
 
 ## Core Responsibilities
 
@@ -5427,80 +5498,74 @@ var TEMPLATES = {
       content: CONTENT
     }
   },
-  "test-code-generator": {
-    playwright: {
+  "test-engineer": {
+    default: {
       frontmatter: FRONTMATTER2,
       content: CONTENT2
     }
   },
-  "test-debugger-fixer": {
-    playwright: {
-      frontmatter: FRONTMATTER3,
-      content: CONTENT3
-    }
-  },
   "team-communicator": {
     local: {
+      frontmatter: FRONTMATTER3,
+      content: CONTENT3
+    },
+    slack: {
       frontmatter: FRONTMATTER4,
       content: CONTENT4
     },
-    slack: {
+    teams: {
       frontmatter: FRONTMATTER5,
       content: CONTENT5
     },
-    teams: {
+    email: {
       frontmatter: FRONTMATTER6,
       content: CONTENT6
-    },
-    email: {
-      frontmatter: FRONTMATTER7,
-      content: CONTENT7
     }
   },
   "documentation-researcher": {
     notion: {
+      frontmatter: FRONTMATTER7,
+      content: CONTENT7
+    },
+    confluence: {
       frontmatter: FRONTMATTER8,
       content: CONTENT8
     },
-    confluence: {
+    jira: {
       frontmatter: FRONTMATTER9,
       content: CONTENT9
-    },
-    jira: {
-      frontmatter: FRONTMATTER10,
-      content: CONTENT10
     }
   },
   "issue-tracker": {
     linear: {
-      frontmatter: FRONTMATTER11,
-      content: CONTENT11
+      frontmatter: FRONTMATTER10,
+      content: CONTENT10
     },
     jira: {
-      frontmatter: FRONTMATTER12,
-      content: CONTENT12
+      frontmatter: FRONTMATTER11,
+      content: CONTENT11
     },
     "jira-server": {
+      frontmatter: FRONTMATTER11,
+      content: CONTENT11
+    },
+    "azure-devops": {
       frontmatter: FRONTMATTER12,
       content: CONTENT12
     },
-    "azure-devops": {
+    notion: {
       frontmatter: FRONTMATTER13,
       content: CONTENT13
     },
-    notion: {
+    slack: {
       frontmatter: FRONTMATTER14,
       content: CONTENT14
-    },
-    slack: {
-      frontmatter: FRONTMATTER15,
-      content: CONTENT15
     }
   },
   "changelog-historian": {
     github: {
-      frontmatter: FRONTMATTER16,
-      content: CONTENT16
+      frontmatter: FRONTMATTER15,
+      content: CONTENT15
     }
   }
 };
@@ -6351,9 +6416,9 @@ var fixTestIssuesStep = {
   category: "execution",
   content: `## Fix Test Issues Automatically
 
-For each test classified as **[TEST ISSUE]**, use the test-debugger-fixer agent to automatically fix the test:
+For each test classified as **[TEST ISSUE]**, use the test-engineer agent to automatically fix the test:
 
-{{INVOKE_TEST_DEBUGGER_FIXER}}
+{{INVOKE_TEST_ENGINEER}}
 
 For each failed test classified as a test issue (not a product bug), provide:
 - Test run timestamp: [from manifest.timestamp]
@@ -6376,7 +6441,7 @@ The agent will:
 7. Repeat up to 3 times if needed (exec-1, exec-2, exec-3)
 8. Report success or escalate as likely product bug
 
-**After test-debugger-fixer completes:**
+**After test-engineer completes:**
 - If fix succeeded: Mark test as fixed, add to "Tests Fixed" list
 - If still failing after 3 attempts: Reclassify as potential product bug
 
@@ -6384,7 +6449,7 @@ The agent will:
 - Maintain list of tests fixed automatically
 - Include fix description (e.g., "Applied selector fix pattern from CLAUDE.md")
 - Note verification status (test now passes)`,
-  invokesSubagents: ["test-debugger-fixer"],
+  invokesSubagents: ["test-engineer"],
   tags: ["execution", "fixing", "automation"]
 };
 
@@ -6755,7 +6820,7 @@ var automateTestCasesStep = {
 
 For each test case marked \`automated: true\`:
 
-{{INVOKE_TEST_CODE_GENERATOR}} with the following context:
+{{INVOKE_TEST_ENGINEER}} with the following context:
 
 "Automate test cases for the focus area: $ARGUMENTS
 
@@ -6788,7 +6853,7 @@ After automation, update the manual test case frontmatter:
 automated: true
 automated_test: tests/specs/feature/test-name.spec.ts
 \`\`\``,
-  invokesSubagents: ["test-code-generator"],
+  invokesSubagents: ["test-engineer"],
   tags: ["generation", "automation"]
 };
 
@@ -7208,6 +7273,29 @@ After test generation completes, verify all artifacts meet quality standards:
   tags: ["maintenance", "validation", "test-artifacts"]
 };
 
+// src/tasks/steps/maintenance/sync-automation-from-feedback.ts
+init_cjs_shims();
+var syncAutomationFromFeedbackStep = {
+  id: "sync-automation-from-feedback",
+  title: "Sync Automation Code with Test Case Changes",
+  category: "maintenance",
+  requiresSubagent: "test-engineer",
+  invokesSubagents: ["test-engineer"],
+  content: `## Sync Automation Code with Test Case Changes
+
+After processing feedback that modified test case files, ensure automation code stays in sync.
+
+**When to apply:** Only if you modified test case markdown files during this task.
+
+For each modified test case:
+1. Read frontmatter \u2014 check for \`automated_test\` field
+2. If exists \u2192 {{INVOKE_TEST_ENGINEER}} to update the spec to match the updated markdown
+3. If not \u2192 skip (no automation to sync)
+
+**CRITICAL:** Both specification (markdown) AND implementation (automation code) MUST stay in sync.`,
+  tags: ["maintenance", "automation", "feedback"]
+};
+
 // src/tasks/steps/index.ts
 init_types();
 var STEP_LIBRARY = {
@@ -7245,7 +7333,8 @@ var STEP_LIBRARY = {
   "generate-final-report": generateFinalReportStep,
   "update-exploration-artifacts": updateExplorationArtifactsStep,
   "cleanup-temp-files": cleanupTempFilesStep,
-  "validate-test-artifacts": validateTestArtifactsStep
+  "validate-test-artifacts": validateTestArtifactsStep,
+  "sync-automation-from-feedback": syncAutomationFromFeedbackStep
 };
 function getStep(id) {
   const step = STEP_LIBRARY[id];
@@ -7351,8 +7440,7 @@ init_cjs_shims();
 var TOOL_STRINGS = {
   "claude-code": {
     INVOKE_BROWSER_AUTOMATION: '**DELEGATE TO SUBAGENT**: Use the Task tool with `subagent_type: "browser-automation"` to delegate test execution.\nThe browser-automation agent will handle all browser automation. DO NOT execute Playwright MCP tools directly.\nInclude the test case path and any specific instructions in the prompt.',
-    INVOKE_TEST_DEBUGGER_FIXER: '**DELEGATE TO SUBAGENT**: Use the Task tool with `subagent_type: "test-debugger-fixer"` to delegate debugging.\nThe agent will analyze failures and fix test code. Include error details and test path in the prompt.',
-    INVOKE_TEST_CODE_GENERATOR: '**DELEGATE TO SUBAGENT**: Use the Task tool with `subagent_type: "test-code-generator"` to delegate code generation.\nThe agent will create automated tests and page objects. Include test case files in the prompt.',
+    INVOKE_TEST_ENGINEER: '**DELEGATE TO SUBAGENT**: Use the Task tool with `subagent_type: "test-engineer"` to delegate test automation work.\nThe agent handles creating, updating, and fixing automated tests. Include test case files and context in the prompt.',
     INVOKE_TEAM_COMMUNICATOR: '**DELEGATE TO SUBAGENT**: Use the Task tool with `subagent_type: "team-communicator"` to send team notifications.\nThe agent will post to Slack/Teams/Email. Include message content and context in the prompt.',
     INLINE_TEAM_COMMUNICATOR: "**TEAM COMMUNICATION**: Read `.claude/agents/team-communicator.md` and follow its instructions to communicate with the team.\nUse the tools and guidelines specified in that file within this context. Do NOT spawn a sub-agent.",
     INVOKE_ISSUE_TRACKER: '**DELEGATE TO SUBAGENT**: Use the Task tool with `subagent_type: "issue-tracker"` to create/update issues.\nThe agent will interact with Jira. Include bug details and classification in the prompt.',
@@ -7361,8 +7449,7 @@ var TOOL_STRINGS = {
   },
   "cursor": {
     INVOKE_BROWSER_AUTOMATION: 'Run the browser-automation agent:\n```bash\ncursor-agent -p "$(cat .cursor/agents/browser-automation.md)" --output-format text\n```',
-    INVOKE_TEST_DEBUGGER_FIXER: 'Run the test-debugger-fixer agent:\n```bash\ncursor-agent -p "$(cat .cursor/agents/test-debugger-fixer.md)" --output-format text\n```',
-    INVOKE_TEST_CODE_GENERATOR: 'Run the test-code-generator agent:\n```bash\ncursor-agent -p "$(cat .cursor/agents/test-code-generator.md)" --output-format text\n```',
+    INVOKE_TEST_ENGINEER: 'Run the test-engineer agent:\n```bash\ncursor-agent -p "$(cat .cursor/agents/test-engineer.md)" --output-format text\n```',
     INVOKE_TEAM_COMMUNICATOR: 'Run the team-communicator agent:\n```bash\ncursor-agent -p "$(cat .cursor/agents/team-communicator.md)" --output-format text\n```',
     INLINE_TEAM_COMMUNICATOR: "**TEAM COMMUNICATION**: Read `.cursor/agents/team-communicator.md` and follow its instructions to communicate with the team.\nUse the tools and guidelines specified in that file within this context.",
     INVOKE_ISSUE_TRACKER: 'Run the issue-tracker agent:\n```bash\ncursor-agent -p "$(cat .cursor/agents/issue-tracker.md)" --output-format text\n```',
@@ -7371,8 +7458,7 @@ var TOOL_STRINGS = {
   },
   "codex": {
     INVOKE_BROWSER_AUTOMATION: 'Run the browser-automation agent:\n```bash\ncodex -p "$(cat .codex/agents/browser-automation.md)"\n```',
-    INVOKE_TEST_DEBUGGER_FIXER: 'Run the test-debugger-fixer agent:\n```bash\ncodex -p "$(cat .codex/agents/test-debugger-fixer.md)"\n```',
-    INVOKE_TEST_CODE_GENERATOR: 'Run the test-code-generator agent:\n```bash\ncodex -p "$(cat .codex/agents/test-code-generator.md)"\n```',
+    INVOKE_TEST_ENGINEER: 'Run the test-engineer agent:\n```bash\ncodex -p "$(cat .codex/agents/test-engineer.md)"\n```',
     INVOKE_TEAM_COMMUNICATOR: 'Run the team-communicator agent:\n```bash\ncodex -p "$(cat .codex/agents/team-communicator.md)"\n```',
     INLINE_TEAM_COMMUNICATOR: "**TEAM COMMUNICATION**: Read `.codex/agents/team-communicator.md` and follow its instructions to communicate with the team.\nUse the tools and guidelines specified in that file within this context.",
     INVOKE_ISSUE_TRACKER: 'Run the issue-tracker agent:\n```bash\ncodex -p "$(cat .codex/agents/issue-tracker.md)"\n```',
@@ -7395,8 +7481,7 @@ function replaceInvocationPlaceholders(content, toolId, isLocal = false) {
   let result = content;
   const keys = [
     "INVOKE_BROWSER_AUTOMATION",
-    "INVOKE_TEST_DEBUGGER_FIXER",
-    "INVOKE_TEST_CODE_GENERATOR",
+    "INVOKE_TEST_ENGINEER",
     "INVOKE_TEAM_COMMUNICATOR",
     "INVOKE_ISSUE_TRACKER",
     "INVOKE_DOCUMENTATION_RESEARCHER",