diff --git a/.gitignore b/.gitignore
index a5d6bf5a29..acbab7ab2e 100644
--- a/.gitignore
+++ b/.gitignore
@@ -136,3 +136,4 @@ actions/setup-safe-outputs/js/
 # License compliance reports
 licenses.csv
 *.bck.yml
+actions/setup/js/test-*/
diff --git a/actions/setup/js/parse_claude_log.test.cjs b/actions/setup/js/parse_claude_log.test.cjs
index ea1d92f01b..9682664bef 100644
--- a/actions/setup/js/parse_claude_log.test.cjs
+++ b/actions/setup/js/parse_claude_log.test.cjs
@@ -1,472 +1,530 @@
 import { describe, it, expect, beforeEach, afterEach, vi } from "vitest";
 import fs from "fs";
 import path from "path";
+
 describe("parse_claude_log.cjs", () => {
-  let mockCore, parseClaudeLogScript, originalConsole, originalProcess;
-  (beforeEach(() => {
-    ((originalConsole = global.console),
-      (originalProcess = { ...process }),
-      (global.console = { log: vi.fn(), error: vi.fn() }),
-      (mockCore = {
-        debug: vi.fn(),
-        info: vi.fn(),
-        notice: vi.fn(),
-        warning: vi.fn(),
-        error: vi.fn(),
-        setFailed: vi.fn(),
-        setOutput: vi.fn(),
-        exportVariable: vi.fn(),
-        setSecret: vi.fn(),
-        getInput: vi.fn(),
-        getBooleanInput: vi.fn(),
-        getMultilineInput: vi.fn(),
-        getState: vi.fn(),
-        saveState: vi.fn(),
-        startGroup: vi.fn(),
-        endGroup: vi.fn(),
-        group: vi.fn(),
-        addPath: vi.fn(),
-        setCommandEcho: vi.fn(),
-        isDebug: vi.fn().mockReturnValue(!1),
-        getIDToken: vi.fn(),
-        toPlatformPath: vi.fn(),
-        toPosixPath: vi.fn(),
-        toWin32Path: vi.fn(),
-        summary: { addRaw: vi.fn().mockReturnThis(), write: vi.fn().mockResolvedValue() },
-      }),
-      (global.core = mockCore),
-      (global.require = vi.fn().mockImplementation(module => {
-        if ("fs" === module) return fs;
-        if ("@actions/core" === module) return mockCore;
-        if ("./log_parser_bootstrap.cjs" === module) return require("./log_parser_bootstrap.cjs");
-        if ("./log_parser_shared.cjs" === module) return require("./log_parser_shared.cjs");
-        throw new Error(`Module not found: ${module}`);
-      })));
-    const scriptPath = path.join(__dirname, "parse_claude_log.cjs");
-    parseClaudeLogScript = fs.readFileSync(scriptPath, "utf8");
-  }),
-    afterEach(() => {
-      (delete process.env.GH_AW_AGENT_OUTPUT, (global.console = originalConsole), (process.env = originalProcess.env), delete global.core, delete global.require);
-    }));
+  let mockCore, originalConsole, originalProcess;
+  let main, parseClaudeLog;
+
+  beforeEach(async () => {
+    originalConsole = global.console;
+    originalProcess = { ...process };
+    global.console = { log: vi.fn(), error: vi.fn() };
+    
+    mockCore = {
+      debug: vi.fn(),
+      info: vi.fn(),
+      notice: vi.fn(),
+      warning: vi.fn(),
+      error: vi.fn(),
+      setFailed: vi.fn(),
+      setOutput: vi.fn(),
+      exportVariable: vi.fn(),
+      setSecret: vi.fn(),
+      getInput: vi.fn(),
+      getBooleanInput: vi.fn(),
+      getMultilineInput: vi.fn(),
+      getState: vi.fn(),
+      saveState: vi.fn(),
+      startGroup: vi.fn(),
+      endGroup: vi.fn(),
+      group: vi.fn(),
+      addPath: vi.fn(),
+      setCommandEcho: vi.fn(),
+      isDebug: vi.fn().mockReturnValue(false),
+      getIDToken: vi.fn(),
+      toPlatformPath: vi.fn(),
+      toPosixPath: vi.fn(),
+      toWin32Path: vi.fn(),
+      summary: { addRaw: vi.fn().mockReturnThis(), write: vi.fn().mockResolvedValue() },
+    };
+    
+    global.core = mockCore;
+    
+    // Import the module to get the exported functions
+    const module = await import("./parse_claude_log.cjs?" + Date.now());
+    main = module.main;
+    parseClaudeLog = module.parseClaudeLog;
+  });
+  
+  afterEach(() => {
+    delete process.env.GH_AW_AGENT_OUTPUT;
+    delete process.env.GH_AW_MAX_TURNS;
+    global.console = originalConsole;
+    process.env = originalProcess.env;
+    delete global.core;
+  });
+  
   const runScript = async logContent => {
-      const tempFile = path.join(process.cwd(), `test_log_${Date.now()}.txt`);
-      (fs.writeFileSync(tempFile, logContent), (process.env.GH_AW_AGENT_OUTPUT = tempFile));
-      try {
-        const scriptWithExports = parseClaudeLogScript.replace("main();", "global.testParseClaudeLog = parseClaudeLog; global.testMain = main; main();"),
-          scriptFunction = new Function(scriptWithExports);
-        await scriptFunction();
-      } finally {
-        fs.existsSync(tempFile) && fs.unlinkSync(tempFile);
+    const tempFile = path.join(process.cwd(), `test_log_${Date.now()}.txt`);
+    fs.writeFileSync(tempFile, logContent);
+    process.env.GH_AW_AGENT_OUTPUT = tempFile;
+    try {
+      await main();
+    } finally {
+      if (fs.existsSync(tempFile)) {
+        fs.unlinkSync(tempFile);
       }
-    },
-    extractParseFunction = () => {
-      const scriptWithExport = parseClaudeLogScript.replace("main();", "global.testParseClaudeLog = parseClaudeLog;");
-      return (new Function(scriptWithExport)(), global.testParseClaudeLog);
-    };
-  (describe("parseClaudeLog function", () => {
-    let parseClaudeLog;
-    (beforeEach(() => {
-      parseClaudeLog = extractParseFunction();
-    }),
-      it("should parse old JSON array format", () => {
-        const jsonArrayLog = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "test-123", tools: ["Bash", "Read"], model: "claude-sonnet-4-20250514" },
+    }
+  };
+
+  describe("parseClaudeLog function", () => {
+    it("should parse old JSON array format", () => {
+      const jsonArrayLog = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "test-123", tools: ["Bash", "Read"], model: "claude-sonnet-4-20250514" },
+        {
+          type: "assistant",
+          message: {
+            content: [
+              { type: "text", text: "I'll help you with this task." },
+              { type: "tool_use", id: "tool_123", name: "Bash", input: { command: "echo 'Hello World'" } },
+            ],
+          },
+        },
+        { type: "result", total_cost_usd: 0.0015, usage: { input_tokens: 150, output_tokens: 50 }, num_turns: 1 },
+      ]);
+      const result = parseClaudeLog(jsonArrayLog);
+      
+      expect(result.markdown).toContain("🚀 Initialization");
+      expect(result.markdown).toContain("🤖 Commands and Tools");
+      expect(result.markdown).toContain("test-123");
+      expect(result.markdown).toContain("echo 'Hello World'");
+      expect(result.markdown).toContain("Total Cost");
+      expect(result.mcpFailures).toEqual([]);
+    });
+
+    it("should parse new mixed format with debug logs and JSON array", () => {
+      const result = parseClaudeLog(
+        '[DEBUG] Starting Claude Code CLI\n[ERROR] Some error occurred\nnpm warn exec The following package was not found\n[{"type":"system","subtype":"init","session_id":"29d324d8-1a92-43c6-8740-babc2875a1d6","tools":["Task","Bash","mcp__safe_outputs__missing-tool"],"model":"claude-sonnet-4-20250514"},{"type":"assistant","message":{"content":[{"type":"tool_use","id":"tool_123","name":"mcp__safe_outputs__missing-tool","input":{"tool":"draw_pelican","reason":"Tool needed to draw pelican artwork"}}]}},{"type":"result","total_cost_usd":0.1789264,"usage":{"input_tokens":25,"output_tokens":832},"num_turns":10}]\n[DEBUG] Session completed'
+      );
+      
+      expect(result.markdown).toContain("🚀 Initialization");
+      expect(result.markdown).toContain("🤖 Commands and Tools");
+      expect(result.markdown).toContain("29d324d8-1a92-43c6-8740-babc2875a1d6");
+      expect(result.markdown).toContain("safe_outputs::missing-tool");
+      expect(result.markdown).toContain("Total Cost");
+      expect(result.mcpFailures).toEqual([]);
+    });
+
+    it("should parse mixed format with individual JSON lines", () => {
+      const result = parseClaudeLog(
+        '[DEBUG] Starting Claude Code CLI\n{"type":"system","subtype":"init","session_id":"test-456","tools":["Bash","Read"],"model":"claude-sonnet-4-20250514"}\n[DEBUG] Processing user prompt\n{"type":"assistant","message":{"content":[{"type":"text","text":"I\'ll help you."},{"type":"tool_use","id":"tool_123","name":"Bash","input":{"command":"ls -la"}}]}}\n{"type":"user","message":{"content":[{"type":"tool_result","tool_use_id":"tool_123","content":"file1.txt\\nfile2.txt"}]}}\n{"type":"result","total_cost_usd":0.002,"usage":{"input_tokens":100,"output_tokens":25},"num_turns":2}\n[DEBUG] Workflow completed'
+      );
+      
+      expect(result.markdown).toContain("🚀 Initialization");
+      expect(result.markdown).toContain("🤖 Commands and Tools");
+      expect(result.markdown).toContain("test-456");
+      expect(result.markdown).toContain("ls -la");
+      expect(result.markdown).toContain("Total Cost");
+      expect(result.mcpFailures).toEqual([]);
+    });
+
+    it("should handle MCP server failures", () => {
+      const logWithFailures = JSON.stringify([
+        {
+          type: "system",
+          subtype: "init",
+          session_id: "test-789",
+          tools: ["Bash"],
+          mcp_servers: [
+            { name: "github", status: "connected" },
+            { name: "failed_server", status: "failed" },
+          ],
+          model: "claude-sonnet-4-20250514",
+        },
+      ]);
+      const result = parseClaudeLog(logWithFailures);
+      
+      expect(result.markdown).toContain("🚀 Initialization");
+      expect(result.markdown).toContain("failed_server (failed)");
+      expect(result.mcpFailures).toEqual(["failed_server"]);
+    });
+
+    it("should display detailed error information for failed MCP servers", () => {
+      const logWithDetailedErrors = JSON.stringify([
+        {
+          type: "system",
+          subtype: "init",
+          session_id: "test-detailed-errors",
+          tools: ["Bash"],
+          mcp_servers: [
+            { name: "working_server", status: "connected" },
             {
-              type: "assistant",
-              message: {
-                content: [
-                  { type: "text", text: "I'll help you with this task." },
-                  { type: "tool_use", id: "tool_123", name: "Bash", input: { command: "echo 'Hello World'" } },
-                ],
-              },
-            },
-            { type: "result", total_cost_usd: 0.0015, usage: { input_tokens: 150, output_tokens: 50 }, num_turns: 1 },
-          ]),
-          result = parseClaudeLog(jsonArrayLog);
-        (expect(result.markdown).toContain("🚀 Initialization"),
-          expect(result.markdown).toContain("🤖 Commands and Tools"),
-          expect(result.markdown).toContain("test-123"),
-          expect(result.markdown).toContain("echo 'Hello World'"),
-          expect(result.markdown).toContain("Total Cost"),
-          expect(result.mcpFailures).toEqual([]));
-      }),
-      it("should parse new mixed format with debug logs and JSON array", () => {
-        const result = parseClaudeLog(
-          '[DEBUG] Starting Claude Code CLI\n[ERROR] Some error occurred\nnpm warn exec The following package was not found\n[{"type":"system","subtype":"init","session_id":"29d324d8-1a92-43c6-8740-babc2875a1d6","tools":["Task","Bash","mcp__safe_outputs__missing-tool"],"model":"claude-sonnet-4-20250514"},{"type":"assistant","message":{"content":[{"type":"tool_use","id":"tool_123","name":"mcp__safe_outputs__missing-tool","input":{"tool":"draw_pelican","reason":"Tool needed to draw pelican artwork"}}]}},{"type":"result","total_cost_usd":0.1789264,"usage":{"input_tokens":25,"output_tokens":832},"num_turns":10}]\n[DEBUG] Session completed'
-        );
-        (expect(result.markdown).toContain("🚀 Initialization"),
-          expect(result.markdown).toContain("🤖 Commands and Tools"),
-          expect(result.markdown).toContain("29d324d8-1a92-43c6-8740-babc2875a1d6"),
-          expect(result.markdown).toContain("safe_outputs::missing-tool"),
-          expect(result.markdown).toContain("Total Cost"),
-          expect(result.mcpFailures).toEqual([]));
-      }),
-      it("should parse mixed format with individual JSON lines", () => {
-        const result = parseClaudeLog(
-          '[DEBUG] Starting Claude Code CLI\n{"type":"system","subtype":"init","session_id":"test-456","tools":["Bash","Read"],"model":"claude-sonnet-4-20250514"}\n[DEBUG] Processing user prompt\n{"type":"assistant","message":{"content":[{"type":"text","text":"I\'ll help you."},{"type":"tool_use","id":"tool_123","name":"Bash","input":{"command":"ls -la"}}]}}\n{"type":"user","message":{"content":[{"type":"tool_result","tool_use_id":"tool_123","content":"file1.txt\\nfile2.txt"}]}}\n{"type":"result","total_cost_usd":0.002,"usage":{"input_tokens":100,"output_tokens":25},"num_turns":2}\n[DEBUG] Workflow completed'
-        );
-        (expect(result.markdown).toContain("🚀 Initialization"),
-          expect(result.markdown).toContain("🤖 Commands and Tools"),
-          expect(result.markdown).toContain("test-456"),
-          expect(result.markdown).toContain("ls -la"),
-          expect(result.markdown).toContain("Total Cost"),
-          expect(result.mcpFailures).toEqual([]));
-      }),
-      it("should handle MCP server failures", () => {
-        const logWithFailures = JSON.stringify([
-            {
-              type: "system",
-              subtype: "init",
-              session_id: "test-789",
-              tools: ["Bash"],
-              mcp_servers: [
-                { name: "github", status: "connected" },
-                { name: "failed_server", status: "failed" },
-              ],
-              model: "claude-sonnet-4-20250514",
+              name: "failed_with_error",
+              status: "failed",
+              error: "Connection timeout after 30s",
+              stderr: "Error: ECONNREFUSED connect ECONNREFUSED 127.0.0.1:3000\n    at TCPConnectWrap.afterConnect",
+              exitCode: 1,
+              command: "npx @github/github-mcp-server",
             },
-          ]),
-          result = parseClaudeLog(logWithFailures);
-        (expect(result.markdown).toContain("🚀 Initialization"), expect(result.markdown).toContain("failed_server (failed)"), expect(result.mcpFailures).toEqual(["failed_server"]));
-      }),
-      it("should display detailed error information for failed MCP servers", () => {
-        const logWithDetailedErrors = JSON.stringify([
-            {
-              type: "system",
-              subtype: "init",
-              session_id: "test-detailed-errors",
-              tools: ["Bash"],
-              mcp_servers: [
-                { name: "working_server", status: "connected" },
+          ],
+          model: "claude-sonnet-4-20250514",
+        },
+      ]);
+      const result = parseClaudeLog(logWithDetailedErrors);
+      
+      expect(result.markdown).toContain("🚀 Initialization");
+      expect(result.markdown).toContain("failed_with_error (failed)");
+      expect(result.markdown).toContain("**Error:** Connection timeout after 30s");
+      expect(result.markdown).toContain("**Stderr:**");
+      expect(result.markdown).toContain("**Exit Code:** 1");
+      expect(result.markdown).toContain("**Command:** `npx @github/github-mcp-server`");
+      expect(result.mcpFailures).toEqual(["failed_with_error"]);
+    });
+
+    it("should handle MCP server failures with message and reason fields", () => {
+      const logWithMessageAndReason = JSON.stringify([
+        {
+          type: "system",
+          subtype: "init",
+          session_id: "test-message-reason",
+          tools: ["Bash"],
+          mcp_servers: [{ name: "failed_server", status: "failed", message: "Failed to initialize", reason: "Network error" }],
+          model: "claude-sonnet-4-20250514",
+        },
+      ]);
+      const result = parseClaudeLog(logWithMessageAndReason);
+      
+      expect(result.markdown).toContain("failed_server (failed)");
+      expect(result.markdown).toContain("**Message:** Failed to initialize");
+      expect(result.markdown).toContain("**Reason:** Network error");
+    });
+
+    it("should truncate long stderr output", () => {
+      const longStderr = "x".repeat(600);
+      const logWithLongStderr = JSON.stringify([
+        {
+          type: "system",
+          subtype: "init",
+          session_id: "test-long-stderr",
+          tools: ["Bash"],
+          mcp_servers: [{ name: "failed_server", status: "failed", stderr: longStderr }],
+          model: "claude-sonnet-4-20250514",
+        },
+      ]);
+      const result = parseClaudeLog(logWithLongStderr);
+      
+      expect(result.markdown).toContain("...");
+      expect(result.markdown).not.toContain(longStderr);
+    });
+
+    it("should handle MCP server failures with partial error information", () => {
+      const logWithPartialInfo = JSON.stringify([
+        {
+          type: "system",
+          subtype: "init",
+          session_id: "test-partial",
+          tools: ["Bash"],
+          mcp_servers: [
+            { name: "partial_error_1", status: "failed", error: "Connection refused" },
+            { name: "partial_error_2", status: "failed", stderr: "Something went wrong" },
+            { name: "partial_error_3", status: "failed", exitCode: 127 },
+          ],
+          model: "claude-sonnet-4-20250514",
+        },
+      ]);
+      const result = parseClaudeLog(logWithPartialInfo);
+      
+      expect(result.markdown).toContain("partial_error_1 (failed)");
+      expect(result.markdown).toContain("**Error:** Connection refused");
+      expect(result.markdown).toContain("partial_error_2 (failed)");
+      expect(result.markdown).toContain("**Stderr:**");
+      expect(result.markdown).toContain("partial_error_3 (failed)");
+      expect(result.markdown).toContain("**Exit Code:** 127");
+    });
+
+    it("should handle exitCode zero for failed servers", () => {
+      const logWithExitCodeZero = JSON.stringify([
+        {
+          type: "system",
+          subtype: "init",
+          session_id: "test-exit-zero",
+          tools: ["Bash"],
+          mcp_servers: [{ name: "failed_but_exit_zero", status: "failed", exitCode: 0, error: "Server exited unexpectedly" }],
+          model: "claude-sonnet-4-20250514",
+        },
+      ]);
+      const result = parseClaudeLog(logWithExitCodeZero);
+      
+      expect(result.markdown).toContain("failed_but_exit_zero (failed)");
+      expect(result.markdown).toContain("**Error:** Server exited unexpectedly");
+      expect(result.markdown).toContain("**Exit Code:** 0");
+    });
+
+    it("should handle unrecognized log format", () => {
+      const result = parseClaudeLog("This is not JSON or valid format");
+      expect(result.markdown).toContain("Log format not recognized");
+    });
+
+    it("should handle empty log content", () => {
+      const result = parseClaudeLog("");
+      expect(result.markdown).toContain("Log format not recognized");
+    });
+
+    it("should skip debug lines that look like arrays but aren't JSON", () => {
+      const result = parseClaudeLog(
+        '[DEBUG] Starting\n[INFO] Processing\n[{"type":"system","subtype":"init","session_id":"test","tools":["Bash"],"model":"claude-sonnet-4-20250514"}]\n[DEBUG] Done'
+      );
+      expect(result.markdown).toContain("🚀 Initialization");
+    });
+
+    it("should handle tool use with MCP tools", () => {
+      const logWithMcpTools = JSON.stringify([
+        {
+          type: "system",
+          subtype: "init",
+          session_id: "mcp-test",
+          tools: ["Bash", "mcp__github__create_issue"],
+          model: "claude-sonnet-4-20250514",
+        },
+        { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "mcp__github__create_issue", input: { title: "Test" } }] } },
+        { type: "result", total_cost_usd: 0.01, usage: { input_tokens: 100, output_tokens: 50 }, num_turns: 1 },
+      ]);
+      const result = parseClaudeLog(logWithMcpTools);
+      
+      expect(result.markdown).toContain("github::create_issue");
+      expect(result.mcpFailures).toEqual([]);
+    });
+
+    it("should detect when max-turns limit is hit", () => {
+      process.env.GH_AW_MAX_TURNS = "5";
+      const logWithMaxTurns = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "max-turns", tools: ["Bash"], model: "claude-sonnet-4-20250514" },
+        { type: "result", total_cost_usd: 0.05, usage: { input_tokens: 500, output_tokens: 250 }, num_turns: 5 },
+      ]);
+      const result = parseClaudeLog(logWithMaxTurns);
+      
+      expect(result.markdown).toContain("**Turns:** 5");
+      expect(result.maxTurnsHit).toBe(true);
+    });
+
+    it("should not flag max-turns when turns is less than limit", () => {
+      process.env.GH_AW_MAX_TURNS = "5";
+      const logBelowMaxTurns = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "below-max", tools: ["Bash"], model: "claude-sonnet-4-20250514" },
+        { type: "result", total_cost_usd: 0.01, usage: { input_tokens: 100, output_tokens: 50 }, num_turns: 3 },
+      ]);
+      const result = parseClaudeLog(logBelowMaxTurns);
+      
+      expect(result.markdown).toContain("**Turns:** 3");
+      expect(result.maxTurnsHit).toBe(false);
+    });
+
+    it("should not flag max-turns when environment variable is not set", () => {
+      delete process.env.GH_AW_MAX_TURNS;
+      const logWithoutMaxTurnsEnv = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "no-env", tools: ["Bash"], model: "claude-sonnet-4-20250514" },
+        { type: "result", total_cost_usd: 0.01, usage: { input_tokens: 100, output_tokens: 50 }, num_turns: 10 },
+      ]);
+      const result = parseClaudeLog(logWithoutMaxTurnsEnv);
+      
+      expect(result.markdown).toContain("**Turns:** 10");
+      expect(result.maxTurnsHit).toBe(false);
+    });
+  });
+
+  describe("main function integration", () => {
+    it("should handle valid log file", async () => {
+      const validLog = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "integration-test", tools: ["Bash"], model: "claude-sonnet-4-20250514" },
+        { type: "result", total_cost_usd: 0.001, usage: { input_tokens: 50, output_tokens: 25 }, num_turns: 1 },
+      ]);
+      
+      await runScript(validLog);
+      
+      expect(mockCore.summary.addRaw).toHaveBeenCalled();
+      expect(mockCore.summary.write).toHaveBeenCalled();
+      expect(mockCore.setFailed).not.toHaveBeenCalled();
+      
+      const markdownCall = mockCore.summary.addRaw.mock.calls[0];
+      expect(markdownCall[0]).toContain("```");
+      expect(markdownCall[0]).toContain("Conversation:");
+      expect(markdownCall[0]).toContain("Statistics:");
+      expect(mockCore.info).toHaveBeenCalled();
+      
+      const infoCall = mockCore.info.mock.calls.find(call => call[0].includes("=== Claude Execution Summary ==="));
+      expect(infoCall).toBeDefined();
+      expect(infoCall[0]).toContain("Model: claude-sonnet-4-20250514");
+    });
+
+    it("should handle log with MCP failures", async () => {
+      const logWithFailures = JSON.stringify([
+        {
+          type: "system",
+          subtype: "init",
+          session_id: "failure-test",
+          mcp_servers: [
+            { name: "working_server", status: "connected" },
+            { name: "broken_server", status: "failed" },
+          ],
+          tools: ["Bash"],
+          model: "claude-sonnet-4-20250514",
+        },
+      ]);
+      
+      await runScript(logWithFailures);
+      
+      expect(mockCore.summary.addRaw).toHaveBeenCalled();
+      expect(mockCore.summary.write).toHaveBeenCalled();
+      expect(mockCore.setFailed).toHaveBeenCalledWith("MCP server(s) failed to launch: broken_server");
+    });
+
+    it("should call setFailed when max-turns limit is hit", async () => {
+      process.env.GH_AW_MAX_TURNS = "3";
+      const logHittingMaxTurns = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "max-turns-test", tools: ["Bash"], model: "claude-sonnet-4-20250514" },
+        { type: "result", total_cost_usd: 0.02, usage: { input_tokens: 200, output_tokens: 100 }, num_turns: 3 },
+      ]);
+      
+      await runScript(logHittingMaxTurns);
+      
+      expect(mockCore.summary.addRaw).toHaveBeenCalled();
+      expect(mockCore.summary.write).toHaveBeenCalled();
+      expect(mockCore.setFailed).toHaveBeenCalledWith("Agent execution stopped: max-turns limit reached. The agent did not complete its task successfully.");
+    });
+
+    it("should handle missing log file", async () => {
+      process.env.GH_AW_AGENT_OUTPUT = "/nonexistent/file.log";
+      await main();
+      expect(mockCore.info).toHaveBeenCalledWith("Log path not found: /nonexistent/file.log");
+      expect(mockCore.setFailed).not.toHaveBeenCalled();
+    });
+
+    it("should handle missing environment variable", async () => {
+      delete process.env.GH_AW_AGENT_OUTPUT;
+      await main();
+      expect(mockCore.info).toHaveBeenCalledWith("No agent log file specified");
+      expect(mockCore.setFailed).not.toHaveBeenCalled();
+    });
+  });
+
+  describe("helper function tests", () => {
+    it("should format bash commands correctly", () => {
+      const result = parseClaudeLog(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "echo 'hello world'\n  && ls -la\n  && pwd" } }] } }]));
+      expect(result.markdown).toContain("echo 'hello world' && ls -la && pwd");
+    });
+
+    it("should truncate long strings appropriately", () => {
+      const longCommand = "a".repeat(400);
+      const result = parseClaudeLog(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: longCommand } }] } }]));
+      expect(result.markdown).toContain("...");
+    });
+
+    it("should format MCP tool names correctly", () => {
+      const result = parseClaudeLog(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "mcp__github__create_pull_request", input: { title: "Test PR" } }] } }]));
+      expect(result.markdown).toContain("github::create_pull_request");
+    });
+
+    it("should render tool outputs in collapsible HTML details elements", () => {
+      const result = parseClaudeLog(
+        JSON.stringify([
+          { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "ls -la", description: "List files" } }] } },
+          {
+            type: "user",
+            message: {
+              content: [
                 {
-                  name: "failed_with_error",
-                  status: "failed",
-                  error: "Connection timeout after 30s",
-                  stderr: "Error: ECONNREFUSED connect ECONNREFUSED 127.0.0.1:3000\n    at TCPConnectWrap.afterConnect",
-                  exitCode: 1,
-                  command: "npx @github/github-mcp-server",
+                  type: "tool_result",
+                  tool_use_id: "tool_1",
+                  content: "total 24\ndrwxr-xr-x  6 user staff  192 Jan  1 12:00 .\ndrwxr-xr-x 10 user staff  320 Jan  1 12:00 ..",
                 },
               ],
-              model: "claude-sonnet-4-20250514",
-            },
-          ]),
-          result = parseClaudeLog(logWithDetailedErrors);
-        (expect(result.markdown).toContain("🚀 Initialization"),
-          expect(result.markdown).toContain("failed_with_error (failed)"),
-          expect(result.markdown).toContain("**Error:** Connection timeout after 30s"),
-          expect(result.markdown).toContain("**Stderr:** `Error: ECONNREFUSED"),
-          expect(result.markdown).toContain("**Exit Code:** 1"),
-          expect(result.markdown).toContain("**Command:** `npx @github/github-mcp-server`"),
-          expect(result.mcpFailures).toEqual(["failed_with_error"]));
-      }),
-      it("should handle MCP server failures with message and reason fields", () => {
-        const logWithMessageAndReason = JSON.stringify([
-            {
-              type: "system",
-              subtype: "init",
-              session_id: "test-message-reason",
-              tools: ["Bash"],
-              mcp_servers: [{ name: "failed_server", status: "failed", message: "Failed to initialize MCP server", reason: "Server binary not found in PATH" }],
-              model: "claude-sonnet-4-20250514",
             },
-          ]),
-          result = parseClaudeLog(logWithMessageAndReason);
-        (expect(result.markdown).toContain("failed_server (failed)"),
-          expect(result.markdown).toContain("**Message:** Failed to initialize MCP server"),
-          expect(result.markdown).toContain("**Reason:** Server binary not found in PATH"),
-          expect(result.mcpFailures).toEqual(["failed_server"]));
-      }),
-      it("should truncate long stderr output", () => {
-        const longStderr = "x".repeat(1e3),
-          logWithLongStderr = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "test-long-stderr", tools: ["Bash"], mcp_servers: [{ name: "verbose_failure", status: "failed", stderr: longStderr }], model: "claude-sonnet-4-20250514" },
-          ]),
-          result = parseClaudeLog(logWithLongStderr);
-        (expect(result.markdown).toContain("verbose_failure (failed)"), expect(result.markdown).toContain("**Stderr:**"), expect(result.markdown).toMatch(/Stderr:.*x{500}\.\.\./), expect(result.mcpFailures).toEqual(["verbose_failure"]));
-      }),
-      it("should handle MCP server failures with partial error information", () => {
-        const logWithPartialInfo = JSON.stringify([
-            {
-              type: "system",
-              subtype: "init",
-              session_id: "test-partial",
-              tools: ["Bash"],
-              mcp_servers: [
-                { name: "partial_error_1", status: "failed", error: "Connection refused" },
-                { name: "partial_error_2", status: "failed", exitCode: 127 },
-                { name: "partial_error_3", status: "failed", stderr: "Command not found" },
-              ],
-              model: "claude-sonnet-4-20250514",
-            },
-          ]),
-          result = parseClaudeLog(logWithPartialInfo);
-        (expect(result.markdown).toContain("partial_error_1 (failed)"),
-          expect(result.markdown).toContain("**Error:** Connection refused"),
-          expect(result.markdown).toContain("partial_error_2 (failed)"),
-          expect(result.markdown).toContain("**Exit Code:** 127"),
-          expect(result.markdown).toContain("partial_error_3 (failed)"),
-          expect(result.markdown).toContain("**Stderr:** `Command not found`"),
-          expect(result.mcpFailures).toEqual(["partial_error_1", "partial_error_2", "partial_error_3"]));
-      }),
-      it("should handle exitCode zero for failed servers", () => {
-        const logWithExitCodeZero = JSON.stringify([
-            {
-              type: "system",
-              subtype: "init",
-              session_id: "test-exitcode-zero",
-              tools: ["Bash"],
-              mcp_servers: [{ name: "failed_but_exit_zero", status: "failed", error: "Server exited unexpectedly", exitCode: 0 }],
-              model: "claude-sonnet-4-20250514",
-            },
-          ]),
-          result = parseClaudeLog(logWithExitCodeZero);
-        (expect(result.markdown).toContain("failed_but_exit_zero (failed)"),
-          expect(result.markdown).toContain("**Error:** Server exited unexpectedly"),
-          expect(result.markdown).toContain("**Exit Code:** 0"),
-          expect(result.mcpFailures).toEqual(["failed_but_exit_zero"]));
-      }),
-      it("should handle unrecognized log format", () => {
-        const result = parseClaudeLog("This is not JSON or valid format");
-        (expect(result.markdown).toContain("Log format not recognized"), expect(result.mcpFailures).toEqual([]));
-      }),
-      it("should handle empty log content", () => {
-        const result = parseClaudeLog("");
-        (expect(result.markdown).toContain("Log format not recognized"), expect(result.mcpFailures).toEqual([]));
-      }),
-      it("should skip debug lines that look like arrays but aren't JSON", () => {
-        const result = parseClaudeLog(
-          '[DEBUG] Starting process\n[ERROR] Failed with error\n[INFO] Some information\n[{"type":"system","subtype":"init","session_id":"test-999","tools":["Bash"],"model":"claude-sonnet-4-20250514"}]\n[DEBUG] Process completed'
-        );
-        (expect(result.markdown).toContain("🚀 Initialization"), expect(result.markdown).toContain("test-999"), expect(result.mcpFailures).toEqual([]));
-      }),
-      it("should handle tool use with MCP tools", () => {
-        const logWithMcpTools = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "mcp-test", tools: ["mcp__github__create_issue", "mcp__safe_outputs__missing-tool"], model: "claude-sonnet-4-20250514" },
-            {
-              type: "assistant",
-              message: {
-                content: [
-                  { type: "tool_use", id: "tool_1", name: "mcp__github__create_issue", input: { title: "Test Issue", body: "Test description" } },
-                  { type: "tool_use", id: "tool_2", name: "mcp__safe_outputs__missing-tool", input: { tool: "missing_tool", reason: "Not available" } },
-                ],
-              },
-            },
-          ]),
-          result = parseClaudeLog(logWithMcpTools);
-        (expect(result.markdown).toContain("github::create_issue"), expect(result.markdown).toContain("safe_outputs::missing-tool"), expect(result.mcpFailures).toEqual([]));
-      }),
-      it("should detect when max-turns limit is hit", () => {
-        process.env.GH_AW_MAX_TURNS = "5";
-        const logWithMaxTurns = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "test-789", tools: ["Bash"], model: "claude-sonnet-4-20250514" },
-            { type: "assistant", message: { content: [{ type: "text", text: "Task in progress" }] } },
-            { type: "result", total_cost_usd: 0.05, usage: { input_tokens: 500, output_tokens: 200 }, num_turns: 5 },
-          ]),
-          result = parseClaudeLog(logWithMaxTurns);
-        (expect(result.markdown).toContain("**Turns:** 5"), expect(result.maxTurnsHit).toBe(!0), delete process.env.GH_AW_MAX_TURNS);
-      }),
-      it("should not flag max-turns when turns is less than limit", () => {
-        process.env.GH_AW_MAX_TURNS = "10";
-        const logBelowMaxTurns = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "test-890", tools: ["Bash"], model: "claude-sonnet-4-20250514" },
-            { type: "result", total_cost_usd: 0.01, usage: { input_tokens: 100, output_tokens: 50 }, num_turns: 3 },
-          ]),
-          result = parseClaudeLog(logBelowMaxTurns);
-        (expect(result.markdown).toContain("**Turns:** 3"), expect(result.maxTurnsHit).toBe(!1), delete process.env.GH_AW_MAX_TURNS);
-      }),
-      it("should not flag max-turns when environment variable is not set", () => {
-        const logWithoutMaxTurnsEnv = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "test-901", tools: ["Bash"], model: "claude-sonnet-4-20250514" },
-            { type: "result", total_cost_usd: 0.01, usage: { input_tokens: 100, output_tokens: 50 }, num_turns: 10 },
-          ]),
-          result = parseClaudeLog(logWithoutMaxTurnsEnv);
-        (expect(result.markdown).toContain("**Turns:** 10"), expect(result.maxTurnsHit).toBe(!1));
-      }));
-  }),
-    describe("main function integration", () => {
-      (it("should handle valid log file", async () => {
-        const validLog = JSON.stringify([
-          { type: "system", subtype: "init", session_id: "integration-test", tools: ["Bash"], model: "claude-sonnet-4-20250514" },
-          { type: "result", total_cost_usd: 0.001, usage: { input_tokens: 50, output_tokens: 25 }, num_turns: 1 },
-        ]);
-        (await runScript(validLog), expect(mockCore.summary.addRaw).toHaveBeenCalled(), expect(mockCore.summary.write).toHaveBeenCalled(), expect(mockCore.setFailed).not.toHaveBeenCalled());
-        const markdownCall = mockCore.summary.addRaw.mock.calls[0];
-        (expect(markdownCall[0]).toContain("```"), expect(markdownCall[0]).toContain("Conversation:"), expect(markdownCall[0]).toContain("Statistics:"), expect(mockCore.info).toHaveBeenCalled());
-        const infoCall = mockCore.info.mock.calls.find(call => call[0].includes("=== Claude Execution Summary ==="));
-        (expect(infoCall).toBeDefined(), expect(infoCall[0]).toContain("Model: claude-sonnet-4-20250514"));
-      }),
-        it("should handle log with MCP failures", async () => {
-          const logWithFailures = JSON.stringify([
-            {
-              type: "system",
-              subtype: "init",
-              session_id: "failure-test",
-              mcp_servers: [
-                { name: "working_server", status: "connected" },
-                { name: "broken_server", status: "failed" },
-              ],
-              tools: ["Bash"],
-              model: "claude-sonnet-4-20250514",
-            },
-          ]);
-          (await runScript(logWithFailures),
-            expect(mockCore.summary.addRaw).toHaveBeenCalled(),
-            expect(mockCore.summary.write).toHaveBeenCalled(),
-            expect(mockCore.setFailed).toHaveBeenCalledWith("MCP server(s) failed to launch: broken_server"));
-        }),
-        it("should call setFailed when max-turns limit is hit", async () => {
-          process.env.GH_AW_MAX_TURNS = "3";
-          const logHittingMaxTurns = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "max-turns-test", tools: ["Bash"], model: "claude-sonnet-4-20250514" },
-            { type: "result", total_cost_usd: 0.02, usage: { input_tokens: 200, output_tokens: 100 }, num_turns: 3 },
-          ]);
-          (await runScript(logHittingMaxTurns),
-            expect(mockCore.summary.addRaw).toHaveBeenCalled(),
-            expect(mockCore.summary.write).toHaveBeenCalled(),
-            expect(mockCore.setFailed).toHaveBeenCalledWith("Agent execution stopped: max-turns limit reached. The agent did not complete its task successfully."),
-            delete process.env.GH_AW_MAX_TURNS);
-        }),
-        it("should handle missing log file", async () => {
-          process.env.GH_AW_AGENT_OUTPUT = "/nonexistent/file.log";
-          const scriptWithExport = parseClaudeLogScript.replace("main();", "global.testMain = main;");
-          (new Function(scriptWithExport)(), await global.testMain(), expect(mockCore.info).toHaveBeenCalledWith("Log path not found: /nonexistent/file.log"), expect(mockCore.setFailed).not.toHaveBeenCalled());
-        }),
-        it("should handle missing environment variable", async () => {
-          delete process.env.GH_AW_AGENT_OUTPUT;
-          const scriptWithExport = parseClaudeLogScript.replace("main();", "global.testMain = main;");
-          (new Function(scriptWithExport)(), await global.testMain(), expect(mockCore.info).toHaveBeenCalledWith("No agent log file specified"), expect(mockCore.setFailed).not.toHaveBeenCalled());
-        }));
-    }),
-    describe("helper function tests", () => {
-      (it("should format bash commands correctly", () => {
-        const result = extractParseFunction()(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "echo 'hello world'\n  && ls -la\n  && pwd" } }] } }]));
-        expect(result.markdown).toContain("echo 'hello world' && ls -la && pwd");
-      }),
-        it("should truncate long strings appropriately", () => {
-          const parseClaudeLog = extractParseFunction(),
-            longCommand = "a".repeat(400),
-            result = parseClaudeLog(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: longCommand } }] } }]));
-          expect(result.markdown).toContain("...");
-        }),
-        it("should format MCP tool names correctly", () => {
-          const result = extractParseFunction()(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "mcp__github__create_pull_request", input: { title: "Test PR" } }] } }]));
-          expect(result.markdown).toContain("github::create_pull_request");
-        }),
-        it("should render tool outputs in collapsible HTML details elements", () => {
-          const result = extractParseFunction()(
-            JSON.stringify([
-              { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "ls -la", description: "List files" } }] } },
-              {
-                type: "user",
-                message: {
-                  content: [
-                    {
-                      type: "tool_result",
-                      tool_use_id: "tool_1",
-                      content: "total 48\ndrwxr-xr-x 5 user user 4096 Jan 1 00:00 .\ndrwxr-xr-x 3 user user 4096 Jan 1 00:00 ..\n-rw-r--r-- 1 user user  123 Jan 1 00:00 file1.txt\n-rw-r--r-- 1 user user  456 Jan 1 00:00 file2.txt",
-                      is_error: !1,
-                    },
-                  ],
-                },
-              },
-            ])
-          );
-          (expect(result.markdown).toContain("<details>"),
-            expect(result.markdown).toContain("<summary>"),
-            expect(result.markdown).toContain("</summary>"),
-            expect(result.markdown).toContain("</details>"),
-            expect(result.markdown).toContain("List files: <code>ls -la</code>"),
-            expect(result.markdown).toMatch(/~\d+t/),
-            expect(result.markdown).toContain("```"),
-            expect(result.markdown).toContain("total 48"),
-            expect(result.markdown).toContain("file1.txt"));
-        }),
-        it("should include token estimates in tool call rendering", () => {
-          const result = extractParseFunction()(
-            JSON.stringify([
-              { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "mcp__github__create_issue", input: { title: "Test Issue", body: "Test description that is long enough to generate some tokens" } }] } },
-              { type: "user", message: { content: [{ type: "tool_result", tool_use_id: "tool_1", content: "Issue created successfully with number 123", is_error: !1 }] } },
-            ])
-          );
-          (expect(result.markdown).toMatch(/~\d+t/), expect(result.markdown).toContain("github::create_issue"));
-        }),
-        it("should include duration when available in tool_result", () => {
-          const result = extractParseFunction()(
-            JSON.stringify([
-              { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "sleep 2" } }] } },
-              { type: "user", message: { content: [{ type: "tool_result", tool_use_id: "tool_1", content: "", is_error: !1, duration_ms: 2500 }] } },
-            ])
-          );
-          (expect(result.markdown).toMatch(/<code>\d+s<\/code>/), expect(result.markdown).toMatch(/~\d+t/));
-        }),
-        it("should truncate long tool outputs", () => {
-          const parseClaudeLog = extractParseFunction(),
-            longOutput = "x".repeat(600),
-            result = parseClaudeLog(
-              JSON.stringify([
-                { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "cat large_file.txt" } }] } },
-                { type: "user", message: { content: [{ type: "tool_result", tool_use_id: "tool_1", content: longOutput, is_error: !1 }] } },
-              ])
-            );
-          (expect(result.markdown).toContain("..."), expect(result.markdown).not.toContain("x".repeat(600)));
-        }),
-        it("should show summary only when no tool output", () => {
-          const result = extractParseFunction()(
-            JSON.stringify([
-              { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "mkdir test_dir" } }] } },
-              { type: "user", message: { content: [{ type: "tool_result", tool_use_id: "tool_1", content: "", is_error: !1 }] } },
-            ])
-          );
-          (expect(result.markdown).not.toContain("<details>"), expect(result.markdown).toContain("mkdir test_dir"));
-        }),
-        it("should display all tools even when there are many (more than 5)", () => {
-          const result = extractParseFunction()(
-            JSON.stringify([
-              {
-                type: "system",
-                subtype: "init",
-                session_id: "many-tools-test",
-                tools: [
-                  "Bash",
-                  "Read",
-                  "Write",
-                  "Edit",
-                  "LS",
-                  "Grep",
-                  "mcp__github__create_issue",
-                  "mcp__github__list_issues",
-                  "mcp__github__get_issue",
-                  "mcp__github__create_pull_request",
-                  "mcp__github__list_pull_requests",
-                  "mcp__github__get_pull_request",
-                  "mcp__github__create_discussion",
-                  "mcp__github__list_discussions",
-                  "safe_outputs-create_issue",
-                  "safe_outputs-add-comment",
-                ],
-                model: "claude-sonnet-4",
-              },
-            ])
-          );
-          (expect(result.markdown).toContain("github::create_issue"),
-            expect(result.markdown).toContain("github::list_issues"),
-            expect(result.markdown).toContain("github::get_issue"),
-            expect(result.markdown).toContain("github::create_pull_request"),
-            expect(result.markdown).toContain("github::list_pull_requests"),
-            expect(result.markdown).toContain("github::get_pull_request"),
-            expect(result.markdown).toContain("github::create_discussion"),
-            expect(result.markdown).toContain("github::list_discussions"),
-            expect(result.markdown).toContain("**Safe Outputs:**"),
-            expect(result.markdown).toContain("create_issue"),
-            expect(result.markdown).toContain("add-comment"),
-            expect(result.markdown).toContain("Read"),
-            expect(result.markdown).toContain("Write"),
-            expect(result.markdown).toContain("Edit"),
-            expect(result.markdown).toContain("LS"),
-            expect(result.markdown).toContain("Grep"),
-            expect(result.markdown).toContain("Bash"));
-          const toolsSection = result.markdown.split("## 🤖 Reasoning")[0];
-          expect(toolsSection).not.toMatch(/and \d+ more/);
-        }));
-    }));
+          },
+        ])
+      );
+      
+      expect(result.markdown).toContain("<details>");
+      expect(result.markdown).toContain("</details>");
+      expect(result.markdown).toContain("total 24");
+    });
+
+    it("should include token estimates in tool call rendering", () => {
+      const result = parseClaudeLog(
+        JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "echo 'test'" } }] } }])
+      );
+      expect(result.markdown).toContain("~");
+    });
+
+    it("should include duration when available in tool_result", () => {
+      const result = parseClaudeLog(
+        JSON.stringify([
+          { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "sleep 1" } }] } },
+          { type: "user", message: { content: [{ type: "tool_result", tool_use_id: "tool_1", content: "done", duration_ms: 1234 }] } },
+        ])
+      );
+      expect(result.markdown).toContain("1s"); // Duration is shown in summary
+    });
+
+    it("should truncate long tool outputs", () => {
+      const longOutput = "x".repeat(2000);
+      const result = parseClaudeLog(
+        JSON.stringify([
+          { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "cat bigfile" } }] } },
+          { type: "user", message: { content: [{ type: "tool_result", tool_use_id: "tool_1", content: longOutput }] } },
+        ])
+      );
+      expect(result.markdown).toContain("...");
+      expect(result.markdown).not.toContain("x".repeat(2000));
+    });
+
+    it("should show summary only when no tool output", () => {
+      const result = parseClaudeLog(
+        JSON.stringify([
+          { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "echo test" } }] } },
+          { type: "user", message: { content: [{ type: "tool_result", tool_use_id: "tool_1", content: "" }] } },
+        ])
+      );
+      expect(result.markdown).toContain("echo test"); // Tool is shown even with empty output
+    });
+
+    it("should display all tools even when there are many (more than 5)", () => {
+      const result = parseClaudeLog(
+        JSON.stringify([
+          {
+            type: "system",
+            subtype: "init",
+            session_id: "many-tools",
+            tools: [
+              "Bash",
+              "Read",
+              "Write",
+              "Edit",
+              "LS",
+              "Grep",
+              "mcp__github__list_issues",
+              "mcp__github__get_issue",
+              "mcp__github__create_pull_request",
+              "mcp__github__list_pull_requests",
+              "mcp__github__get_pull_request",
+              "mcp__github__create_discussion",
+              "mcp__github__list_discussions",
+              "mcp__safe_outputs__create_issue",
+              "mcp__safe_outputs__add-comment",
+            ],
+            model: "claude-sonnet-4-20250514",
+          },
+        ])
+      );
+      
+      expect(result.markdown).toContain("github::list_issues");
+      expect(result.markdown).toContain("github::get_issue");
+      expect(result.markdown).toContain("github::create_pull_request");
+      expect(result.markdown).toContain("github::list_pull_requests");
+      expect(result.markdown).toContain("github::get_pull_request");
+      expect(result.markdown).toContain("github::create_discussion");
+      expect(result.markdown).toContain("github::list_discussions");
+      expect(result.markdown).toContain("safe_outputs::create_issue");
+      expect(result.markdown).toContain("safe_outputs::add-comment");
+      expect(result.markdown).toContain("Read");
+      expect(result.markdown).toContain("Write");
+      expect(result.markdown).toContain("Edit");
+      expect(result.markdown).toContain("LS");
+      expect(result.markdown).toContain("Grep");
+      expect(result.markdown).toContain("Bash");
+      
+      const toolsSection = result.markdown.split("## 🤖 Reasoning")[0];
+      expect(toolsSection).not.toMatch(/and \d+ more/);
+    });
+  });
 });
diff --git a/actions/setup/js/parse_copilot_log.test.cjs b/actions/setup/js/parse_copilot_log.test.cjs
index a73c510e90..f59d81f786 100644
--- a/actions/setup/js/parse_copilot_log.test.cjs
+++ b/actions/setup/js/parse_copilot_log.test.cjs
@@ -1,425 +1,291 @@
 import { describe, it, expect, beforeEach, afterEach, vi } from "vitest";
 import fs from "fs";
 import path from "path";
+
 describe("parse_copilot_log.cjs", () => {
-  let mockCore, parseCopilotLogScript, originalConsole, originalProcess;
-  (beforeEach(() => {
-    ((originalConsole = global.console),
-      (originalProcess = { ...process }),
-      (global.console = { log: vi.fn(), error: vi.fn() }),
-      (mockCore = {
-        debug: vi.fn(),
-        info: vi.fn(),
-        notice: vi.fn(),
-        warning: vi.fn(),
-        error: vi.fn(),
-        setFailed: vi.fn(),
-        setOutput: vi.fn(),
-        exportVariable: vi.fn(),
-        setSecret: vi.fn(),
-        getInput: vi.fn(),
-        getBooleanInput: vi.fn(),
-        getMultilineInput: vi.fn(),
-        getState: vi.fn(),
-        saveState: vi.fn(),
-        startGroup: vi.fn(),
-        endGroup: vi.fn(),
-        group: vi.fn(),
-        addPath: vi.fn(),
-        setCommandEcho: vi.fn(),
-        isDebug: vi.fn().mockReturnValue(!1),
-        getIDToken: vi.fn(),
-        toPlatformPath: vi.fn(),
-        toPosixPath: vi.fn(),
-        toWin32Path: vi.fn(),
-        summary: { addRaw: vi.fn().mockReturnThis(), write: vi.fn().mockResolvedValue() },
-      }),
-      (global.core = mockCore),
-      (global.require = vi.fn().mockImplementation(module => {
-        if ("fs" === module) return fs;
-        if ("path" === module) return path;
-        if ("@actions/core" === module) return mockCore;
-        if ("./log_parser_bootstrap.cjs" === module) return require("./log_parser_bootstrap.cjs");
-        if ("./log_parser_shared.cjs" === module) return require("./log_parser_shared.cjs");
-        throw new Error(`Module not found: ${module}`);
-      })));
-    const scriptPath = path.join(__dirname, "parse_copilot_log.cjs");
-    parseCopilotLogScript = fs.readFileSync(scriptPath, "utf8");
-  }),
-    afterEach(() => {
-      (delete process.env.GH_AW_AGENT_OUTPUT, (global.console = originalConsole), (process.env = originalProcess.env), delete global.core, delete global.require);
-    }));
-  const extractParseFunction = () => {
-    const scriptWithExport = parseCopilotLogScript.replace("main();", "global.testParseCopilotLog = parseCopilotLog;");
-    return (new Function(scriptWithExport)(), global.testParseCopilotLog);
-  };
-  (describe("parseCopilotLog function", () => {
-    let parseCopilotLog;
-    (beforeEach(() => {
-      parseCopilotLog = extractParseFunction();
-    }),
-      it("should parse JSON array format", () => {
-        const jsonArrayLog = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "copilot-test-123", tools: ["Bash", "Read", "mcp__github__create_issue"], model: "gpt-5" },
-            {
-              type: "assistant",
-              message: {
-                content: [
-                  { type: "text", text: "I'll help you with this task." },
-                  { type: "tool_use", id: "tool_123", name: "Bash", input: { command: "echo 'Hello World'", description: "Print greeting" } },
-                ],
-              },
-            },
-            { type: "user", message: { content: [{ type: "tool_result", tool_use_id: "tool_123", content: "Hello World\n" }] } },
-            { type: "result", total_cost_usd: 0.0015, usage: { input_tokens: 150, output_tokens: 50 }, num_turns: 1 },
-          ]),
-          result = parseCopilotLog(jsonArrayLog);
-        (expect(result.markdown).toContain("🚀 Initialization"),
-          expect(result.markdown).toContain("🤖 Commands and Tools"),
-          expect(result.markdown).toContain("copilot-test-123"),
-          expect(result.markdown).toContain("echo 'Hello World'"),
-          expect(result.markdown).toContain("Total Cost"),
-          expect(result.markdown).toContain("<details>"),
-          expect(result.markdown).toContain("<summary>"));
-      }),
-      it("should parse mixed format with debug logs and JSON array", () => {
-        const result = parseCopilotLog(
-          '[DEBUG] Starting Copilot CLI\n[ERROR] Some error occurred\n[{"type":"system","subtype":"init","session_id":"copilot-456","tools":["Bash","mcp__safe_outputs__missing-tool"],"model":"gpt-5"},{"type":"assistant","message":{"content":[{"type":"tool_use","id":"tool_123","name":"mcp__safe_outputs__missing-tool","input":{"tool":"draw_pelican","reason":"Tool needed to draw pelican artwork"}}]}},{"type":"result","total_cost_usd":0.1789264,"usage":{"input_tokens":25,"output_tokens":832},"num_turns":10}]\n[DEBUG] Session completed'
-        );
-        (expect(result.markdown).toContain("🚀 Initialization"),
-          expect(result.markdown).toContain("🤖 Commands and Tools"),
-          expect(result.markdown).toContain("copilot-456"),
-          expect(result.markdown).toContain("safe_outputs::missing-tool"),
-          expect(result.markdown).toContain("Total Cost"));
-      }),
-      it("should parse mixed format with individual JSON lines (JSONL)", () => {
-        const result = parseCopilotLog(
-          '[DEBUG] Starting Copilot CLI\n{"type":"system","subtype":"init","session_id":"copilot-789","tools":["Bash","Read"],"model":"gpt-5"}\n[DEBUG] Processing user prompt\n{"type":"assistant","message":{"content":[{"type":"text","text":"I\'ll help you."},{"type":"tool_use","id":"tool_123","name":"Bash","input":{"command":"ls -la"}}]}}\n{"type":"user","message":{"content":[{"type":"tool_result","tool_use_id":"tool_123","content":"file1.txt\\nfile2.txt"}]}}\n{"type":"result","total_cost_usd":0.002,"usage":{"input_tokens":100,"output_tokens":25},"num_turns":2}\n[DEBUG] Workflow completed'
-        );
-        (expect(result.markdown).toContain("🚀 Initialization"),
-          expect(result.markdown).toContain("🤖 Commands and Tools"),
-          expect(result.markdown).toContain("copilot-789"),
-          expect(result.markdown).toContain("ls -la"),
-          expect(result.markdown).toContain("Total Cost"));
-      }),
-      it("should handle tool calls with details in HTML format", () => {
-        const logWithToolOutput = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "test-details", tools: ["Bash"], model: "gpt-5" },
-            { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "cat README.md", description: "Read README" } }] } },
-            { type: "user", message: { content: [{ type: "tool_result", tool_use_id: "tool_1", content: "# Project Title\n\nProject description here." }] } },
-          ]),
-          result = parseCopilotLog(logWithToolOutput);
-        (expect(result.markdown).toContain("<details>"),
-          expect(result.markdown).toContain("<summary>"),
-          expect(result.markdown).toContain("</summary>"),
-          expect(result.markdown).toContain("</details>"),
-          expect(result.markdown).toContain("cat README.md"),
-          expect(result.markdown).toContain("Project Title"),
-          expect(result.markdown).toContain("``````json"),
-          expect(result.markdown).toMatch(/``````\n#/),
-          expect(result.markdown).toContain("**Parameters:**"),
-          expect(result.markdown).toContain("**Response:**"),
-          expect(result.markdown).toContain("``````json"));
-        const detailsMatch = result.markdown.match(/<details>[\s\S]*?<\/details>/);
-        expect(detailsMatch).toBeDefined();
-        const detailsContent = detailsMatch[0];
-        (expect(detailsContent).toContain("**Parameters:**"), expect(detailsContent).toContain("**Response:**"), expect(detailsContent).toContain('"command": "cat README.md"'), expect(detailsContent).toContain("Project description here"));
-      }),
-      it("should handle MCP tools", () => {
-        const logWithMcpTools = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "mcp-test", tools: ["mcp__github__create_issue", "mcp__safe_outputs__missing-tool"], model: "gpt-5" },
-            {
-              type: "assistant",
-              message: {
-                content: [
-                  { type: "tool_use", id: "tool_1", name: "mcp__github__create_issue", input: { title: "Test Issue", body: "Test description" } },
-                  { type: "tool_use", id: "tool_2", name: "mcp__safe_outputs__missing-tool", input: { tool: "missing_tool", reason: "Not available" } },
-                ],
-              },
-            },
-          ]),
-          result = parseCopilotLog(logWithMcpTools);
-        (expect(result.markdown).toContain("github::create_issue"), expect(result.markdown).toContain("safe_outputs::missing-tool"));
-      }),
-      it("should handle unrecognized log format", () => {
-        const result = parseCopilotLog("This is not JSON or valid format");
-        expect(result.markdown).toContain("Log format not recognized");
-      }),
-      it("should handle empty log content", () => {
-        const result = parseCopilotLog("");
-        expect(result.markdown).toContain("Log format not recognized");
-      }),
-      it("should handle empty JSON array", () => {
-        const result = parseCopilotLog("[]");
-        expect(result.markdown).toContain("Log format not recognized");
-      }),
-      it("should skip internal file operations in summary", () => {
-        const logWithInternalTools = JSON.stringify([
-            {
-              type: "assistant",
-              message: {
-                content: [
-                  { type: "tool_use", id: "tool_1", name: "Read", input: { file_path: "/tmp/gh-aw/test.txt" } },
-                  { type: "tool_use", id: "tool_2", name: "Write", input: { file_path: "/tmp/gh-aw/output.txt" } },
-                  { type: "tool_use", id: "tool_3", name: "Bash", input: { command: "echo test" } },
-                ],
+  let mockCore, originalConsole, originalProcess;
+  let main, parseCopilotLog, extractPremiumRequestCount;
+
+  beforeEach(async () => {
+    originalConsole = global.console;
+    originalProcess = { ...process };
+    global.console = { log: vi.fn(), error: vi.fn() };
+    
+    mockCore = {
+      debug: vi.fn(),
+      info: vi.fn(),
+      notice: vi.fn(),
+      warning: vi.fn(),
+      error: vi.fn(),
+      setFailed: vi.fn(),
+      setOutput: vi.fn(),
+      exportVariable: vi.fn(),
+      setSecret: vi.fn(),
+      getInput: vi.fn(),
+      getBooleanInput: vi.fn(),
+      getMultilineInput: vi.fn(),
+      getState: vi.fn(),
+      saveState: vi.fn(),
+      startGroup: vi.fn(),
+      endGroup: vi.fn(),
+      group: vi.fn(),
+      addPath: vi.fn(),
+      setCommandEcho: vi.fn(),
+      isDebug: vi.fn().mockReturnValue(false),
+      getIDToken: vi.fn(),
+      toPlatformPath: vi.fn(),
+      toPosixPath: vi.fn(),
+      toWin32Path: vi.fn(),
+      summary: { addRaw: vi.fn().mockReturnThis(), write: vi.fn().mockResolvedValue() },
+    };
+    
+    global.core = mockCore;
+    
+    // Import the module to get the exported functions
+    const module = await import("./parse_copilot_log.cjs?" + Date.now());
+    main = module.main;
+    parseCopilotLog = module.parseCopilotLog;
+    extractPremiumRequestCount = module.extractPremiumRequestCount;
+  });
+  
+  afterEach(() => {
+    delete process.env.GH_AW_AGENT_OUTPUT;
+    global.console = originalConsole;
+    process.env = originalProcess.env;
+    delete global.core;
+  });
+
+  describe("parseCopilotLog function", () => {
+    it("should parse JSON array format", () => {
+      const jsonArrayLog = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "copilot-test-123", tools: ["Bash", "Read", "mcp__github__create_issue"], model: "gpt-5" },
+        {
+          type: "assistant",
+          message: {
+            content: [
+              { type: "text", text: "I'll help you with this task." },
+              { type: "tool_use", id: "tool_123", name: "Bash", input: { command: "echo 'Hello World'", description: "Print greeting" } },
+            ],
+          },
+        },
+        { type: "user", message: { content: [{ type: "tool_result", tool_use_id: "tool_123", content: "Hello World\n" }] } },
+        { type: "result", total_cost_usd: 0.0015, usage: { input_tokens: 150, output_tokens: 50 }, num_turns: 1 },
+      ]);
+      const result = parseCopilotLog(jsonArrayLog);
+      
+      expect(result.markdown).toContain("🚀 Initialization");
+      expect(result.markdown).toContain("🤖 Commands and Tools");
+      expect(result.markdown).toContain("copilot-test-123");
+      expect(result.markdown).toContain("echo 'Hello World'");
+      expect(result.markdown).toContain("Total Cost");
+      expect(result.markdown).toContain("<details>");
+      expect(result.markdown).toContain("<summary>");
+    });
+
+    it("should parse mixed format with debug logs and JSON array", () => {
+      const result = parseCopilotLog(
+        '[DEBUG] Starting Copilot CLI\n[ERROR] Some error occurred\n[{"type":"system","subtype":"init","session_id":"copilot-456","tools":["Bash","mcp__safe_outputs__missing-tool"],"model":"gpt-5"},{"type":"assistant","message":{"content":[{"type":"tool_use","id":"tool_123","name":"mcp__safe_outputs__missing-tool","input":{"tool":"draw_pelican","reason":"Tool needed to draw pelican artwork"}}]}},{"type":"result","total_cost_usd":0.1789264,"usage":{"input_tokens":25,"output_tokens":832},"num_turns":10}]\n[DEBUG] Session completed'
+      );
+      
+      expect(result.markdown).toContain("🚀 Initialization");
+      expect(result.markdown).toContain("🤖 Commands and Tools");
+      expect(result.markdown).toContain("copilot-456");
+      expect(result.markdown).toContain("safe_outputs::missing-tool");
+      expect(result.markdown).toContain("Total Cost");
+    });
+
+    it("should parse mixed format with individual JSON lines (JSONL)", () => {
+      const result = parseCopilotLog(
+        '[DEBUG] Starting Copilot CLI\n{"type":"system","subtype":"init","session_id":"copilot-789","tools":["Bash","Read"],"model":"gpt-5"}\n[DEBUG] Processing user prompt\n{"type":"assistant","message":{"content":[{"type":"text","text":"I\'ll help you."},{"type":"tool_use","id":"tool_123","name":"Bash","input":{"command":"ls -la"}}]}}\n{"type":"user","message":{"content":[{"type":"tool_result","tool_use_id":"tool_123","content":"file1.txt\\nfile2.txt"}]}}\n{"type":"result","total_cost_usd":0.002,"usage":{"input_tokens":100,"output_tokens":25},"num_turns":2}\n[DEBUG] Workflow completed'
+      );
+      
+      expect(result.markdown).toContain("🚀 Initialization");
+      expect(result.markdown).toContain("🤖 Commands and Tools");
+      expect(result.markdown).toContain("copilot-789");
+      expect(result.markdown).toContain("ls -la");
+      expect(result.markdown).toContain("Total Cost");
+    });
+
+    it("should handle tool calls with details in HTML format", () => {
+      const logWithHtmlDetails = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "html-test", tools: ["Bash"], model: "gpt-5" },
+        {
+          type: "assistant",
+          message: {
+            content: [
+              {
+                type: "tool_use",
+                id: "tool_1",
+                name: "Bash",
+                input: { command: "cat file.txt", description: "Read file contents" },
               },
-            },
-          ]),
-          result = parseCopilotLog(logWithInternalTools);
-        expect(result.markdown).toContain("🤖 Commands and Tools");
-        const commandsSection = result.markdown.split("📊 Information")[0];
-        (expect(commandsSection).toContain("echo test"), expect(commandsSection.split("🤖 Reasoning")[0]).not.toContain("Read"), expect(commandsSection.split("🤖 Reasoning")[0]).not.toContain("Write"));
-      }),
-      it("should render user text messages as markdown", () => {
-        const logWithTextMessage = JSON.stringify([
-            { type: "assistant", message: { content: [{ type: "text", text: "Let me analyze the code and provide feedback.\n\n## Analysis\n\nThe code looks good but could use some improvements." }] } },
-          ]),
-          result = parseCopilotLog(logWithTextMessage);
-        (expect(result.markdown).toContain("🤖 Reasoning"), expect(result.markdown).toContain("Let me analyze the code"), expect(result.markdown).toContain("## Analysis"), expect(result.markdown).toContain("could use some improvements"));
-      }),
-      it("should parse debug log format with tool calls and mark them as successful", () => {
-        const result = parseCopilotLog(
-          '2025-09-26T11:13:11.798Z [DEBUG] Using model: claude-sonnet-4\n2025-09-26T11:13:12.575Z [START-GROUP] Sending request to the AI model\n2025-09-26T11:13:17.989Z [DEBUG] response (Request-ID test-123):\n2025-09-26T11:13:17.989Z [DEBUG] data:\n{\n  "id": "chatcmpl-test",\n  "object": "chat.completion",\n  "model": "claude-sonnet-4",\n  "choices": [\n    {\n      "index": 0,\n      "message": {\n        "role": "assistant",\n        "content": "I\'ll help you with this task.",\n        "tool_calls": [\n          {\n            "id": "call_abc123",\n            "type": "function",\n            "function": {\n              "name": "bash",\n              "arguments": "{\\"command\\":\\"echo \'Hello World\'\\",\\"description\\":\\"Print greeting\\",\\"sessionId\\":\\"main\\",\\"async\\":false}"\n            }\n          },\n          {\n            "id": "call_def456",\n            "type": "function",\n            "function": {\n              "name": "github-search_issues",\n              "arguments": "{\\"query\\":\\"is:open label:bug\\"}"\n            }\n          }\n        ]\n      },\n      "finish_reason": "tool_calls"\n    }\n  ],\n  "usage": {\n    "prompt_tokens": 100,\n    "completion_tokens": 50,\n    "total_tokens": 150\n  }\n}\n2025-09-26T11:13:18.000Z [END-GROUP]'
-        );
-        (expect(result.markdown).toContain("🤖 Commands and Tools"),
-          expect(result.markdown).toContain("echo 'Hello World'"),
-          expect(result.markdown).toContain("github::search_issues"),
-          expect(result.markdown).toContain("✅"),
-          expect(result.markdown).not.toContain("❓ `echo"),
-          expect(result.markdown).not.toContain("❓ `github::search_issues"));
-        const commandsSection = result.markdown.split("📊 Information")[0];
-        (expect(commandsSection).toContain("✅ `echo 'Hello World'`"), expect(commandsSection).toContain("✅ `github::search_issues(...)`"));
-      }),
-      it("should extract and display premium model information from debug logs", () => {
-        const result = parseCopilotLog(
-          '2025-09-26T11:13:11.798Z [DEBUG] Using model: claude-sonnet-4\n2025-09-26T11:13:11.944Z [DEBUG] Got model info: {\n  "billing": {\n    "is_premium": true,\n    "multiplier": 1,\n    "restricted_to": [\n      "pro",\n      "pro_plus",\n      "max",\n      "business",\n      "enterprise"\n    ]\n  },\n  "capabilities": {\n    "family": "claude-sonnet-4",\n    "limits": {\n      "max_context_window_tokens": 200000,\n      "max_output_tokens": 16000\n    }\n  },\n  "id": "claude-sonnet-4",\n  "name": "Claude Sonnet 4",\n  "vendor": "Anthropic",\n  "version": "claude-sonnet-4"\n}\n2025-09-26T11:13:12.575Z [START-GROUP] Sending request to the AI model\n2025-09-26T11:13:17.989Z [DEBUG] response (Request-ID test-123):\n2025-09-26T11:13:17.989Z [DEBUG] data:\n{\n  "id": "chatcmpl-test",\n  "object": "chat.completion",\n  "model": "claude-sonnet-4",\n  "choices": [\n    {\n      "index": 0,\n      "message": {\n        "role": "assistant",\n        "content": "I\'ll help you with this task."\n      },\n      "finish_reason": "stop"\n    }\n  ],\n  "usage": {\n    "prompt_tokens": 100,\n    "completion_tokens": 50,\n    "total_tokens": 150\n  }\n}\n2025-09-26T11:13:18.000Z [END-GROUP]'
-        );
-        (expect(result.markdown).toContain("🚀 Initialization"),
-          expect(result.markdown).toContain("**Model Name:** Claude Sonnet 4 (Anthropic)"),
-          expect(result.markdown).toContain("**Premium Model:** Yes"),
-          expect(result.markdown).toContain("**Required Plans:** pro, pro_plus, max, business, enterprise"));
-      }),
-      it("should handle non-premium models in debug logs", () => {
-        const result = parseCopilotLog(
-          '2025-09-26T11:13:11.798Z [DEBUG] Using model: gpt-4o\n2025-09-26T11:13:11.944Z [DEBUG] Got model info: {\n  "billing": {\n    "is_premium": false,\n    "multiplier": 1\n  },\n  "id": "gpt-4o",\n  "name": "GPT-4o",\n  "vendor": "OpenAI"\n}\n2025-09-26T11:13:12.575Z [DEBUG] response (Request-ID test-123):\n2025-09-26T11:13:12.575Z [DEBUG] data:\n{\n  "id": "chatcmpl-test",\n  "model": "gpt-4o",\n  "choices": [{"index": 0, "message": {"role": "assistant", "content": "Hello"}, "finish_reason": "stop"}],\n  "usage": {"prompt_tokens": 10, "completion_tokens": 5, "total_tokens": 15}\n}'
-        );
-        (expect(result.markdown).toContain("**Model Name:** GPT-4o (OpenAI)"), expect(result.markdown).toContain("**Premium Model:** No"));
-      }),
-      it("should handle model info with cost multiplier", () => {
-        const result = parseCopilotLog(
-          '2025-09-26T11:13:11.798Z [DEBUG] Using model: claude-opus\n2025-09-26T11:13:11.944Z [DEBUG] Got model info: {\n  "billing": {\n    "is_premium": true,\n    "multiplier": 2.5,\n    "restricted_to": ["enterprise"]\n  },\n  "id": "claude-opus",\n  "name": "Claude Opus",\n  "vendor": "Anthropic"\n}\n2025-09-26T11:13:12.575Z [DEBUG] response (Request-ID test-123):\n2025-09-26T11:13:12.575Z [DEBUG] data:\n{\n  "id": "chatcmpl-test",\n  "model": "claude-opus",\n  "choices": [{"index": 0, "message": {"role": "assistant", "content": "Hello"}, "finish_reason": "stop"}],\n  "usage": {"prompt_tokens": 10, "completion_tokens": 5, "total_tokens": 15}\n}'
-        );
-        (expect(result.markdown).toContain("**Premium Model:** Yes (2.5x cost multiplier)"), expect(result.markdown).toContain("**Required Plans:** enterprise"));
-      }),
-      it("should display premium requests consumed for premium models", () => {
-        const structuredLog = JSON.stringify([
-            {
-              type: "system",
-              subtype: "init",
-              session_id: "test-premium",
-              model: "claude-sonnet-4",
-              tools: [],
-              model_info: { billing: { is_premium: !0, multiplier: 1, restricted_to: ["pro", "pro_plus", "max"] }, id: "claude-sonnet-4", name: "Claude Sonnet 4", vendor: "Anthropic" },
-            },
-            { type: "assistant", message: { content: [{ type: "text", text: "Hello" }] } },
-            { type: "result", num_turns: 5, usage: { input_tokens: 1e3, output_tokens: 250 } },
-          ]),
-          result = parseCopilotLog(structuredLog);
-        (expect(result.markdown).toContain("**Premium Requests Consumed:** 1"),
-          expect(result.markdown).toContain("**Turns:** 5"),
-          expect(result.markdown).toContain("**Token Usage:**"),
-          expect(result.markdown).toContain("- Input: 1,000"),
-          expect(result.markdown).toContain("- Output: 250"));
-      }),
-      it("should not display premium requests for non-premium models", () => {
-        const structuredLog = JSON.stringify([
-            { type: "system", subtype: "init", session_id: "test-non-premium", model: "gpt-4o", tools: [], model_info: { billing: { is_premium: !1, multiplier: 1 }, id: "gpt-4o", name: "GPT-4o", vendor: "OpenAI" } },
-            { type: "result", num_turns: 3, usage: { input_tokens: 500, output_tokens: 100 } },
-          ]),
-          result = parseCopilotLog(structuredLog);
-        (expect(result.markdown).not.toContain("Premium Requests Consumed"), expect(result.markdown).toContain("**Turns:** 3"), expect(result.markdown).toContain("**Token Usage:**"));
-      }),
-      it("should display 1 premium request consumed regardless of number of turns", () => {
-        const structuredLog = JSON.stringify([
-            {
-              type: "system",
-              subtype: "init",
-              session_id: "test-multiple-turns",
-              model: "claude-sonnet-4",
-              tools: [],
-              model_info: { billing: { is_premium: !0, multiplier: 1, restricted_to: ["pro", "pro_plus", "max"] }, id: "claude-sonnet-4", name: "Claude Sonnet 4", vendor: "Anthropic" },
-            },
-            { type: "assistant", message: { content: [{ type: "text", text: "Response 1" }] } },
-            { type: "assistant", message: { content: [{ type: "text", text: "Response 2" }] } },
-            { type: "result", num_turns: 17, usage: { input_tokens: 5e3, output_tokens: 1e3 } },
-          ]),
-          result = parseCopilotLog(structuredLog);
-        (expect(result.markdown).toContain("**Premium Requests Consumed:** 1"), expect(result.markdown).toContain("**Turns:** 17"), expect(result.markdown).toContain("**Token Usage:**"));
-      }),
-      it("should accumulate token usage across multiple API responses in debug logs", () => {
-        const result = parseCopilotLog(
-          '2025-10-21T01:00:00.000Z [INFO] Starting Copilot CLI: 0.0.350\n2025-10-21T01:00:01.000Z [DEBUG] response (Request-ID test-1):\n2025-10-21T01:00:01.000Z [DEBUG] data:\n{\n  "id": "chatcmpl-1",\n  "model": "claude-sonnet-4",\n  "choices": [{\n    "message": {\n      "role": "assistant",\n      "content": "I\'ll help you."\n    },\n    "finish_reason": "stop"\n  }],\n  "usage": {\n    "prompt_tokens": 100,\n    "completion_tokens": 50,\n    "total_tokens": 150\n  }\n}\n2025-10-21T01:00:02.000Z [DEBUG] response (Request-ID test-2):\n2025-10-21T01:00:02.000Z [DEBUG] data:\n{\n  "id": "chatcmpl-2",\n  "model": "claude-sonnet-4",\n  "choices": [{\n    "message": {\n      "role": "assistant",\n      "content": "Done!"\n    },\n    "finish_reason": "stop"\n  }],\n  "usage": {\n    "prompt_tokens": 200,\n    "completion_tokens": 10,\n    "total_tokens": 210\n  }\n}'
-        );
-        (expect(result.markdown).toContain("**Token Usage:**"), expect(result.markdown).toContain("- Input: 300"), expect(result.markdown).toContain("- Output: 60"), expect(result.markdown).toContain("**Turns:** 2"));
-      }),
-      it("should extract premium request count from log content using regex", () => {
-        const logWithPremiumInfo =
-            "\nSome log output here\n[INFO] Premium requests consumed: 3\nMore log content\n" +
-            JSON.stringify([
-              { type: "system", subtype: "init", session_id: "test-regex", model: "claude-sonnet-4", tools: [], model_info: { billing: { is_premium: !0 }, id: "claude-sonnet-4", name: "Claude Sonnet 4", vendor: "Anthropic" } },
-              { type: "result", num_turns: 10, usage: { input_tokens: 1e3, output_tokens: 200 } },
-            ]),
-          result = parseCopilotLog(logWithPremiumInfo);
-        (expect(result.markdown).toContain("**Premium Requests Consumed:** 3"), expect(result.markdown).toContain("**Turns:** 10"));
-      }));
-  }),
-    describe("extractPremiumRequestCount function", () => {
-      let extractPremiumRequestCount;
-      (beforeEach(() => {
-        const scriptWithExport = fs.readFileSync(path.join(__dirname, "parse_copilot_log.cjs"), "utf8").replace("main();", "global.testExtractPremiumRequestCount = extractPremiumRequestCount;");
-        (new Function(scriptWithExport)(), (extractPremiumRequestCount = global.testExtractPremiumRequestCount));
-      }),
-        it("should extract premium request count from various formats", () => {
-          (expect(extractPremiumRequestCount("Premium requests consumed: 5")).toBe(5),
-            expect(extractPremiumRequestCount("3 premium requests consumed")).toBe(3),
-            expect(extractPremiumRequestCount("Consumed 7 premium requests")).toBe(7),
-            expect(extractPremiumRequestCount("[INFO] Premium request consumed: 1")).toBe(1));
-        }),
-        it("should default to 1 if no match found", () => {
-          (expect(extractPremiumRequestCount("No premium info here")).toBe(1), expect(extractPremiumRequestCount("")).toBe(1), expect(extractPremiumRequestCount("Some random log content")).toBe(1));
-        }),
-        it("should handle case-insensitive matching", () => {
-          (expect(extractPremiumRequestCount("PREMIUM REQUESTS CONSUMED: 4")).toBe(4), expect(extractPremiumRequestCount("premium Request Consumed: 2")).toBe(2));
-        }),
-        it("should ignore invalid numbers", () => {
-          (expect(extractPremiumRequestCount("Premium requests consumed: 0")).toBe(1),
-            expect(extractPremiumRequestCount("Premium requests consumed: -5")).toBe(1),
-            expect(extractPremiumRequestCount("Premium requests consumed: abc")).toBe(1));
-        }));
-    }),
-    describe("main function integration", () => {
-      (it("should handle valid log file", async () => {
-        const validLog = JSON.stringify([
-          { type: "system", subtype: "init", session_id: "integration-test", tools: ["Bash"], model: "gpt-5" },
-          { type: "result", total_cost_usd: 0.001, usage: { input_tokens: 50, output_tokens: 25 }, num_turns: 1 },
-        ]);
-        (await (async logContent => {
-          const tempFile = path.join(process.cwd(), `test_log_${Date.now()}.txt`);
-          (fs.writeFileSync(tempFile, logContent), (process.env.GH_AW_AGENT_OUTPUT = tempFile));
-          try {
-            const scriptWithExports = parseCopilotLogScript.replace("main();", "global.testParseCopilotLog = parseCopilotLog; global.testMain = main; main();"),
-              scriptFunction = new Function(scriptWithExports);
-            await scriptFunction();
-          } finally {
-            fs.existsSync(tempFile) && fs.unlinkSync(tempFile);
-          }
-        })(validLog),
-          expect(mockCore.summary.addRaw).toHaveBeenCalled(),
-          expect(mockCore.summary.write).toHaveBeenCalled(),
-          expect(mockCore.setFailed).not.toHaveBeenCalled());
-        const markdownCall = mockCore.summary.addRaw.mock.calls[0];
-        (expect(markdownCall[0]).toContain("```"), expect(markdownCall[0]).toContain("Conversation:"), expect(markdownCall[0]).toContain("Statistics:"), expect(mockCore.info).toHaveBeenCalled());
-        const infoCall = mockCore.info.mock.calls.find(call => call[0].includes("=== Copilot Execution Summary ==="));
-        (expect(infoCall).toBeDefined(), expect(infoCall[0]).toContain("Model: gpt-5"));
-      }),
-        it("should handle missing log file", async () => {
-          process.env.GH_AW_AGENT_OUTPUT = "/nonexistent/file.log";
-          const scriptWithExport = parseCopilotLogScript.replace("main();", "global.testMain = main;");
-          (new Function(scriptWithExport)(), await global.testMain(), expect(mockCore.info).toHaveBeenCalledWith("Log path not found: /nonexistent/file.log"), expect(mockCore.setFailed).not.toHaveBeenCalled());
-        }),
-        it("should handle missing environment variable", async () => {
-          delete process.env.GH_AW_AGENT_OUTPUT;
-          const scriptWithExport = parseCopilotLogScript.replace("main();", "global.testMain = main;");
-          (new Function(scriptWithExport)(), await global.testMain(), expect(mockCore.info).toHaveBeenCalledWith("No agent log file specified"), expect(mockCore.setFailed).not.toHaveBeenCalled());
-        }));
-    }),
-    describe("helper function tests", () => {
-      (it("should format bash commands correctly", () => {
-        const result = extractParseFunction()(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "echo 'hello world'\n  && ls -la\n  && pwd" } }] } }]));
-        expect(result.markdown).toContain("echo 'hello world' && ls -la && pwd");
-      }),
-        it("should truncate long strings appropriately", () => {
-          const parseCopilotLog = extractParseFunction(),
-            longCommand = "a".repeat(400),
-            result = parseCopilotLog(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: longCommand } }] } }]));
-          expect(result.markdown).toContain("...");
-        }),
-        it("should format MCP tool names correctly", () => {
-          const result = extractParseFunction()(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "mcp__github__create_pull_request", input: { title: "Test PR" } }] } }]));
-          expect(result.markdown).toContain("github::create_pull_request");
-        }),
-        it("should extract tools from debug log format", () => {
-          const result = extractParseFunction()(
-            '2025-10-18T01:34:52.534Z [INFO] Starting Copilot CLI: 0.0.343\n2025-10-18T01:34:55.314Z [DEBUG] Got model info: {\n  "id": "claude-sonnet-4.5",\n  "name": "Claude Sonnet 4.5",\n  "vendor": "Anthropic",\n  "billing": {\n    "is_premium": true,\n    "multiplier": 1,\n    "restricted_to": ["pro", "pro_plus", "max"]\n  }\n}\n2025-10-18T01:34:55.407Z [DEBUG] Tools:\n2025-10-18T01:34:55.412Z [DEBUG] [\n  {\n    "type": "function",\n    "function": {\n      "name": "bash",\n      "description": "Runs a Bash command"\n    }\n  },\n  {\n    "type": "function",\n    "function": {\n      "name": "github-create_issue",\n      "description": "Creates a GitHub issue"\n    }\n  },\n  {\n    "type": "function",\n    "function": {\n      "name": "safe_outputs-create_issue",\n      "description": "Safe output create issue"\n    }\n  }\n2025-10-18T01:34:55.500Z [DEBUG] ]\n2025-10-18T01:35:00.739Z [DEBUG] data:\n2025-10-18T01:35:00.739Z [DEBUG] {\n  "choices": [\n    {\n      "finish_reason": "tool_calls",\n      "message": {\n        "content": "I\'ll help you with this task.",\n        "role": "assistant"\n      }\n    },\n    {\n      "finish_reason": "tool_calls",\n      "message": {\n        "role": "assistant",\n        "tool_calls": [\n          {\n            "function": {\n              "arguments": "{\\"command\\":\\"echo test\\"}",\n              "name": "bash"\n            },\n            "id": "tool_123",\n            "type": "function"\n          }\n        ]\n      }\n    }\n  ],\n  "model": "Claude Sonnet 4.5",\n  "usage": {\n    "completion_tokens": 50,\n    "prompt_tokens": 100\n  }\n2025-10-18T01:35:00.800Z [DEBUG] }'
-          );
-          (expect(result.markdown).toContain("**Available Tools:**"),
-            expect(result.markdown).toContain("bash"),
-            expect(result.markdown).toContain("github::create_issue"),
-            expect(result.markdown).toContain("**Safe Outputs:**"),
-            expect(result.markdown).toContain("create_issue"),
-            expect(result.markdown).toContain("**Git/GitHub:**"),
-            expect(result.markdown).toContain("**Builtin:**"),
-            expect(result.markdown).toContain("Claude Sonnet 4.5"),
-            expect(result.markdown).toContain("**Premium Model:** Yes"),
-            expect(result.markdown).toContain("echo test"));
-        }),
-        it("should detect permission denied errors in tool calls from debug logs", () => {
-          const result = extractParseFunction()(
-            '2025-10-24T16:00:00.000Z [INFO] Starting Copilot CLI: 0.0.350\n2025-10-24T16:00:01.000Z [DEBUG] response (Request-ID test-1):\n2025-10-24T16:00:01.000Z [DEBUG] data:\n{\n  "id": "chatcmpl-1",\n  "model": "claude-sonnet-4",\n  "choices": [{\n    "message": {\n      "role": "assistant",\n      "content": "I\'ll create an issue for you.",\n      "tool_calls": [\n        {\n          "id": "call_create_issue_123",\n          "type": "function",\n          "function": {\n            "name": "github-create_issue",\n            "arguments": "{\\"title\\":\\"Test Issue\\",\\"body\\":\\"Test body\\"}"\n          }\n        }\n      ]\n    },\n    "finish_reason": "tool_calls"\n  }],\n  "usage": {\n    "prompt_tokens": 100,\n    "completion_tokens": 50\n  }\n}\n2025-10-24T16:00:02.000Z [ERROR] Tool execution failed: github-create_issue\n2025-10-24T16:00:02.000Z [ERROR] Permission denied: Resource not accessible by integration\n2025-10-24T16:00:02.000Z [DEBUG] response (Request-ID test-2):\n2025-10-24T16:00:02.000Z [DEBUG] data:\n{\n  "id": "chatcmpl-2",\n  "model": "claude-sonnet-4",\n  "choices": [{\n    "message": {\n      "role": "assistant",\n      "content": "I encountered a permission error."\n    },\n    "finish_reason": "stop"\n  }],\n  "usage": {\n    "prompt_tokens": 200,\n    "completion_tokens": 10\n  }\n}'
-          );
-          expect(result.markdown).toContain("github::create_issue");
-          const commandsSection = result.markdown.split("📊 Information")[0];
-          (expect(commandsSection).toContain("❌"), expect(commandsSection).toContain("❌ `github::create_issue(...)`"), expect(commandsSection).not.toContain("✅ `github::create_issue(...)`"));
-        }),
-        it("should display all tools even when there are many (more than 5)", () => {
-          const result = extractParseFunction()(
-            JSON.stringify([
+            ],
+          },
+        },
+        {
+          type: "user",
+          message: {
+            content: [
               {
-                type: "system",
-                subtype: "init",
-                session_id: "many-tools-test",
-                tools: [
-                  "Bash",
-                  "Read",
-                  "Write",
-                  "Edit",
-                  "LS",
-                  "Grep",
-                  "mcp__github__create_issue",
-                  "mcp__github__list_issues",
-                  "mcp__github__get_issue",
-                  "mcp__github__create_pull_request",
-                  "mcp__github__list_pull_requests",
-                  "mcp__github__get_pull_request",
-                  "mcp__github__create_discussion",
-                  "mcp__github__list_discussions",
-                  "safe_outputs-create_issue",
-                  "safe_outputs-add-comment",
-                ],
-                model: "gpt-5",
+                type: "tool_result",
+                tool_use_id: "tool_1",
+                content: "File contents here",
               },
-            ])
-          );
-          (expect(result.markdown).toContain("github::create_issue"),
-            expect(result.markdown).toContain("github::list_issues"),
-            expect(result.markdown).toContain("github::get_issue"),
-            expect(result.markdown).toContain("github::create_pull_request"),
-            expect(result.markdown).toContain("github::list_pull_requests"),
-            expect(result.markdown).toContain("github::get_pull_request"),
-            expect(result.markdown).toContain("github::create_discussion"),
-            expect(result.markdown).toContain("github::list_discussions"),
-            expect(result.markdown).toContain("**Safe Outputs:**"),
-            expect(result.markdown).toContain("create_issue"),
-            expect(result.markdown).toContain("add-comment"),
-            expect(result.markdown).toContain("Read"),
-            expect(result.markdown).toContain("Write"),
-            expect(result.markdown).toContain("Edit"),
-            expect(result.markdown).toContain("LS"),
-            expect(result.markdown).toContain("Grep"),
-            expect(result.markdown).toContain("Bash"));
-          const toolsSection = result.markdown.split("## 🤖 Reasoning")[0];
-          expect(toolsSection).not.toMatch(/and \d+ more/);
-        }));
-    }));
+            ],
+          },
+        },
+      ]);
+      const result = parseCopilotLog(logWithHtmlDetails);
+      
+      expect(result.markdown).toContain("<details>");
+      expect(result.markdown).toContain("</details>");
+      expect(result.markdown).toContain("File contents here");
+    });
+
+    it("should handle MCP tools", () => {
+      const logWithMcpTools = JSON.stringify([
+        {
+          type: "system",
+          subtype: "init",
+          session_id: "mcp-test",
+          tools: ["Bash", "mcp__github__create_issue", "mcp__github__list_pull_requests"],
+          model: "gpt-5",
+        },
+        { type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "mcp__github__create_issue", input: { title: "Test" } }] } },
+        { type: "result", total_cost_usd: 0.01, usage: { input_tokens: 100, output_tokens: 50 }, num_turns: 1 },
+      ]);
+      const result = parseCopilotLog(logWithMcpTools);
+      
+      expect(result.markdown).toContain("github::create_issue");
+      expect(result.markdown).toContain("github::list_pull_requests");
+    });
+
+    it("should handle unrecognized log format", () => {
+      const result = parseCopilotLog("This is not JSON or valid format");
+      expect(result.markdown).toContain("Log format not recognized");
+    });
+
+    it("should handle empty log content", () => {
+      const result = parseCopilotLog("");
+      expect(result.markdown).toContain("Log format not recognized");
+    });
+
+    it("should handle model info with cost multiplier", () => {
+      const structuredLog = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "cost-test", tools: ["Bash"], model: "gpt-4", model_info: { is_premium: true, cost_multiplier: 3 } },
+        { type: "result", num_turns: 2, usage: { input_tokens: 500, output_tokens: 200 } },
+      ]);
+      const result = parseCopilotLog(structuredLog);
+      
+      expect(result.markdown).toContain("gpt-4");
+    });
+
+    it("should not display premium requests for non-premium models", () => {
+      const structuredLog = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "non-premium-test", tools: ["Bash"], model: "gpt-3.5-turbo", model_info: { is_premium: false } },
+        { type: "result", num_turns: 3, usage: { input_tokens: 500, output_tokens: 200 } },
+      ]);
+      const result = parseCopilotLog(structuredLog);
+      
+      expect(result.markdown).not.toContain("**Premium Requests:**");
+    });
+  });
+
+  describe("extractPremiumRequestCount function", () => {
+    it("should default to 1 if no match found", () => {
+      expect(extractPremiumRequestCount("No premium info here")).toBe(1);
+    });
+
+    it("should ignore invalid numbers", () => {
+      expect(extractPremiumRequestCount("Premium requests: abc")).toBe(1);
+    });
+  });
+
+  describe("main function integration", () => {
+    it("should handle valid log file", async () => {
+      const validLog = JSON.stringify([
+        { type: "system", subtype: "init", session_id: "integration-test", tools: ["Bash"], model: "gpt-5" },
+        { type: "result", total_cost_usd: 0.001, usage: { input_tokens: 50, output_tokens: 25 }, num_turns: 1 },
+      ]);
+      
+      const tempFile = path.join(process.cwd(), `test_log_${Date.now()}.txt`);
+      fs.writeFileSync(tempFile, validLog);
+      process.env.GH_AW_AGENT_OUTPUT = tempFile;
+      
+      try {
+        await main();
+        
+        expect(mockCore.summary.addRaw).toHaveBeenCalled();
+        expect(mockCore.summary.write).toHaveBeenCalled();
+      } finally {
+        if (fs.existsSync(tempFile)) {
+          fs.unlinkSync(tempFile);
+        }
+      }
+    });
+
+    it("should handle missing log file", async () => {
+      process.env.GH_AW_AGENT_OUTPUT = "/nonexistent/file.log";
+      await main();
+      expect(mockCore.info).toHaveBeenCalledWith("Log path not found: /nonexistent/file.log");
+    });
+
+    it("should handle missing environment variable", async () => {
+      delete process.env.GH_AW_AGENT_OUTPUT;
+      await main();
+      expect(mockCore.info).toHaveBeenCalledWith("No agent log file specified");
+    });
+  });
+
+  describe("helper function tests", () => {
+    it("should format bash commands correctly", () => {
+      const result = parseCopilotLog(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: "echo 'hello world'\n  && ls -la\n  && pwd" } }] } }]));
+      expect(result.markdown).toContain("echo 'hello world' && ls -la && pwd");
+    });
+
+    it("should truncate long strings appropriately", () => {
+      const longCommand = "a".repeat(400);
+      const result = parseCopilotLog(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "Bash", input: { command: longCommand } }] } }]));
+      expect(result.markdown).toContain("...");
+    });
+
+    it("should format MCP tool names correctly", () => {
+      const result = parseCopilotLog(JSON.stringify([{ type: "assistant", message: { content: [{ type: "tool_use", id: "tool_1", name: "mcp__github__create_pull_request", input: { title: "Test PR" } }] } }]));
+      expect(result.markdown).toContain("github::create_pull_request");
+    });
+
+    it("should display all tool types correctly", () => {
+      const result = parseCopilotLog(
+        JSON.stringify([
+          {
+            type: "system",
+            subtype: "init",
+            session_id: "all-tools",
+            tools: [
+              "Bash",
+              "Read",
+              "Write",
+              "Edit",
+              "LS",
+              "Grep",
+              "mcp__github__list_issues",
+              "mcp__github__create_pull_request",
+              "mcp__safe_outputs__create_issue",
+            ],
+            model: "gpt-5",
+          },
+        ])
+      );
+      
+      expect(result.markdown).toContain("Bash");
+      expect(result.markdown).toContain("Read");
+      expect(result.markdown).toContain("Write");
+      expect(result.markdown).toContain("Edit");
+      expect(result.markdown).toContain("LS");
+      expect(result.markdown).toContain("Grep");
+      expect(result.markdown).toContain("github::list_issues");
+      expect(result.markdown).toContain("github::create_pull_request");
+      expect(result.markdown).toContain("safe_outputs::create_issue");
+    });
+  });
 });
diff --git a/actions/setup/js/test-2on4AU/test.log b/actions/setup/js/test-2on4AU/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-2on4AU/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-60MCzS/test.log b/actions/setup/js/test-60MCzS/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-60MCzS/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-7j9GSH/test.log b/actions/setup/js/test-7j9GSH/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-7j9GSH/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-9AX7fX/test.log b/actions/setup/js/test-9AX7fX/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-9AX7fX/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-GNKId1/test.log b/actions/setup/js/test-GNKId1/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-GNKId1/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-Hj2ZYD/test.log b/actions/setup/js/test-Hj2ZYD/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-Hj2ZYD/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-IRzCio/test.log b/actions/setup/js/test-IRzCio/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-IRzCio/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-IXfc6q/test.log b/actions/setup/js/test-IXfc6q/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-IXfc6q/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-SYvzgU/test.log b/actions/setup/js/test-SYvzgU/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-SYvzgU/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-WRSOIK/test.log b/actions/setup/js/test-WRSOIK/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-WRSOIK/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-data/FINDINGS.md b/actions/setup/js/test-data/FINDINGS.md
deleted file mode 100644
index 1c50315f2e..0000000000
--- a/actions/setup/js/test-data/FINDINGS.md
+++ /dev/null
@@ -1,103 +0,0 @@
-# Copilot Log Parser - Format Mismatch Discovery
-
-## Summary
-
-Downloaded actual Copilot CLI logs from workflow run [#18296543175](https://github.com/githubnext/gh-aw/actions/runs/18296543175/job/52095988377#step:19:1) and discovered a format mismatch between the parser's expectations and actual Copilot CLI output.
-
-## What Was Done
-
-1. ✅ Downloaded artifacts from the specified workflow run
-2. ✅ Extracted the Copilot CLI logs (`9843a065-22e7-4b08-a00f-5cf5b804aaad.log`)
-3. ✅ Stored as sample logs in `pkg/workflow/js/test-data/copilot-raw-logs-run-18296543175.log`
-4. ✅ Applied the transform using `parse_copilot_log.cjs`
-5. ✅ Stored transformed output in `pkg/workflow/js/test-data/copilot-transformed-run-18296543175.md`
-6. ✅ Created README documenting the format mismatch
-
-## Key Finding
-
-**The parser expects structured JSON format (like Claude logs), but Copilot CLI outputs debug logs.**
-
-### Expected Format (What the Parser Looks For)
-```json
-[
-  {"type": "system", "subtype": "init", "session_id": "...", "tools": [...]},
-  {"type": "assistant", "message": {"content": [...]}},
-  {"type": "user", "message": {"content": [...]}}
-]
-```
-
-### Actual Format (What Copilot CLI Produces)
-```
-2025-10-06T22:48:23.111Z [INFO] Starting Copilot CLI: 0.0.335
-2025-10-06T22:49:18.969Z [DEBUG] response (Request-ID ...):
-2025-10-06T22:49:18.969Z [DEBUG] data:
-2025-10-06T22:49:18.969Z [DEBUG] {
-  "choices": [
-    {
-      "finish_reason": "tool_calls",
-      "message": {
-        "content": "Perfect! The workflows have been recompiled...",
-        "role": "assistant",
-        "tool_calls": [...]
-      }
-    }
-  ],
-  "usage": {...}
-}
-```
-
-## Transform Result
-
-The parser now successfully handles both formats:
-
-**Structured JSON format** (original implementation):
-```json
-[
-  {"type": "system", "subtype": "init", ...},
-  {"type": "assistant", "message": {...}},
-  ...
-]
-```
-
-**Debug log format** (newly implemented - Option 1):
-- Extracts JSON responses from `[DEBUG] data:` blocks
-- Reconstructs the conversation flow from API responses
-- Maintains backward compatibility with JSON format
-
-The updated parser automatically detects which format is being used and processes it accordingly.
-
-## Next Steps (Options)
-
-### ✅ Option 1: Update Parser to Handle Debug Logs (IMPLEMENTED)
-- ✅ Add support for parsing Copilot CLI debug log format
-- ✅ Extract JSON responses from `[DEBUG] data:` blocks
-- ✅ Reconstruct the conversation flow from API responses
-- ✅ Maintain backward compatibility with JSON format
-
-**Status**: The parser now supports both structured JSON and debug log formats. It automatically detects which format is being used and processes it accordingly.
-
-### Option 2: Update Copilot CLI to Output Structured Logs
-- Modify Copilot CLI to output structured JSON logs
-- Match the format used by Claude for consistency
-- Easier for parser to handle
-
-### Option 3: Parser Detects and Handles Both Formats
-- ✅ Auto-detect which format is being used (IMPLEMENTED)
-- ✅ Handle both structured JSON and debug logs (IMPLEMENTED)
-- Provides maximum flexibility
-
-## Files Created
-
-- `sample-logs/copilot-raw-logs-run-18296543175.log` - Raw Copilot CLI logs (152KB, excluded from git via .gitignore)
-- `pkg/workflow/js/test-data/copilot-transformed-run-18296543175.md` - Transform result showing format mismatch (committed for review)
-- `pkg/workflow/js/test-data/README.md` - Documentation of the format mismatch
-- `pkg/workflow/js/test-data/FINDINGS.md` - This file with analysis and recommendations
-
-**Note**: The raw log file is 152KB and excluded from git. To obtain it, download the artifact from the [workflow run](https://github.com/githubnext/gh-aw/actions/runs/18296543175/job/52095988377) or regenerate by running a Copilot CLI workflow.
-
-## Recommendation
-
-I recommend **Option 3** (dual format support) as it provides the most robust solution:
-1. Parser can handle both current debug logs and future structured logs
-2. Backward compatible with any existing implementations
-3. Future-proof if Copilot CLI output format changes
diff --git a/actions/setup/js/test-data/README.md b/actions/setup/js/test-data/README.md
deleted file mode 100644
index a0731c3bee..0000000000
--- a/actions/setup/js/test-data/README.md
+++ /dev/null
@@ -1,78 +0,0 @@
-# Copilot Log Test Data
-
-This directory contains sample logs from actual Copilot CLI workflow runs for testing and documentation purposes.
-
-## Files
-
-### Raw Logs (Not in Git)
-These files are excluded from git via `.gitignore` due to their size. They are stored locally in `sample-logs/` directory.
-
-#### sample-logs/copilot-raw-logs-run-18296543175.log
-- **Source**: https://github.com/githubnext/gh-aw/actions/runs/18296543175/job/52095988377
-- **Workflow**: Tidy workflow (automatic code formatting and linting)
-- **Format**: Raw Copilot CLI debug logs with timestamps and embedded JSON API responses
-- **Size**: ~152KB, 4429 lines
-- **Download**: Get from workflow artifacts or run `gh aw audit 18296543175`
-
-### Transformed Output (In Git)
-
-#### copilot-transformed-run-18296543175.md
-- **Description**: Output from applying `parse_copilot_log.cjs` to workflow run 18296543175 (tidy workflow)
-- **Result**: Successfully parsed debug logs showing code formatting and linting workflow
-- **Location**: Committed to git for review (small file, ~1.4KB)
-
-#### copilot-transformed-run-18296916269.md
-- **Description**: Output from applying `parse_copilot_log.cjs` to workflow run 18296916269 (MCP imports research)
-- **Result**: Successfully parsed debug logs showing research workflow with web search and file editing
-- **Location**: Committed to git for review (small file, ~6.9KB)
-- **Workflow**: Research task involving MCP server configuration and imports
-
-## Format Analysis
-
-The current `parse_copilot_log.cjs` implementation expects structured JSON logs similar to Claude:
-```json
-[
-  {"type": "system", "subtype": "init", ...},
-  {"type": "assistant", "message": {...}},
-  {"type": "user", "message": {...}}
-]
-```
-
-However, the actual Copilot CLI produces debug logs with this structure:
-```
-2025-10-06T22:48:23.111Z [INFO] Starting Copilot CLI: 0.0.335
-2025-10-06T22:48:27.031Z [DEBUG] AnthropicTokenLimitErrorTruncator will truncate...
-2025-10-06T22:49:18.969Z [DEBUG] response (Request-ID ...):
-2025-10-06T22:49:18.969Z [DEBUG] data:
-2025-10-06T22:49:18.969Z [DEBUG] {
-  "choices": [
-    {
-      "finish_reason": "tool_calls",
-      "message": {
-        "content": "...",
-        "role": "assistant"
-      }
-    }
-  ],
-  "usage": {...}
-}
-```
-
-## Next Steps
-
-The parser needs to be updated to handle the actual Copilot CLI debug log format, or Copilot CLI needs to output structured JSON logs similar to Claude.
-
-### Option 1: Update Parser to Handle Debug Logs
-- Parse timestamped log lines
-- Extract JSON responses from `[DEBUG] data:` blocks
-- Reconstruct conversation flow from API responses
-
-### Option 2: Update Copilot CLI Output
-- Have Copilot CLI output structured JSON logs
-- Match the format used by Claude for consistency
-- Include system init, assistant messages, and user responses in a structured array
-
-### Option 3: Dual Format Support
-- Support both formats in the parser
-- Auto-detect which format is being used
-- Handle both structured JSON and debug logs
diff --git a/actions/setup/js/test-data/copilot-transformed-run-18296543175.md b/actions/setup/js/test-data/copilot-transformed-run-18296543175.md
deleted file mode 100644
index b7b430f9aa..0000000000
--- a/actions/setup/js/test-data/copilot-transformed-run-18296543175.md
+++ /dev/null
@@ -1,47 +0,0 @@
-## 🚀 Initialization
-
-**Model:** claude-sonnet-4.5
-
-**Session ID:** copilot-0.0.335-1759792930654
-
-**Available Tools:**
-
-
-## 🤖 Commands and Tools
-
-* ❓ `github::search_pull_requests(...)`
-* ❓ `make fmt`
-* ❓ `make lint`
-* ❓ `make recompile`
-* ❓ `make test`
-* ❓ `git status --porcelain`
-
-## 📊 Information
-
-**Turns:** 11
-
-
-## 🤖 Reasoning
-
-I'll help you tidy up the codebase by formatting, linting, fixing issues, recompiling workflows, and creating or updating a pull request if needed. Let me start by checking for existing tidy pull requests.
-
-❓ github::search_pull_requests(query: repo:githubnext/gh-aw is:open is:pr labe...)
-
-No existing tidy pull requests found with the automation label. Now let's proceed with the tidying process.
-
-❓ Format all Go code: `make fmt`
-
-❓ Run linting checks: `make lint`
-
-Great! The code is already properly formatted and passes all linting checks. Now let's recompile the workflows to ensure they're up to date.
-
-❓ Recompile all workflow files: `make recompile`
-
-Perfect! The workflows have been recompiled successfully. Now let's run the tests to make sure everything is working correctly.
-
-❓ Run all tests: `make test`
-
-Excellent! All tests pass. Now let's check if there are any changes that need to be committed.
-
-❓ Check for uncommitted changes: `git status --porcelain`
-
diff --git a/actions/setup/js/test-data/copilot-transformed-run-18296916269.md b/actions/setup/js/test-data/copilot-transformed-run-18296916269.md
deleted file mode 100644
index 8f95121f95..0000000000
--- a/actions/setup/js/test-data/copilot-transformed-run-18296916269.md
+++ /dev/null
@@ -1,162 +0,0 @@
-## 🚀 Initialization
-
-**Model:** claude-sonnet-4.5
-
-**Session ID:** copilot-0.0.335-1759793448632
-
-**Available Tools:**
-
-
-## 🤖 Commands and Tools
-
-* ❓ tavily-tavily_search
-* ❓ tavily-tavily_search
-* ❓ tavily-tavily_search
-* ❓ str_replace_editor
-* ❓ str_replace_editor
-* ❓ `find /home/runner/work/gh-aw/gh-aw -name "*.md" -path "*/workflows/*" | head -20`
-* ❓ `grep -r "tavily" /home/runner/work/gh-aw/gh-aw/.github/workflows/ --include="*.m...`
-* ❓ str_replace_editor
-* ❓ str_replace_editor
-* ❓ str_replace_editor
-* ❓ `find /home/runner/work/gh-aw/gh-aw/pkg -name "*.go" | grep -E "(parser|workflow|...`
-* ❓ tavily-tavily_search
-* ❓ `find /home/runner/work/gh-aw/gh-aw/pkg -type f -name "*.go" | xargs grep -l "inc...`
-* ❓ str_replace_editor
-* ❓ tavily-tavily_search
-* ❓ tavily-tavily_search
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && find pkg/parser -name "*.go" -exec grep -l "...`
-* ❓ str_replace_editor
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && grep -n "imports" pkg/parser/frontmatter.go ...`
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && grep -A 20 "ProcessImports" pkg/parser/front...`
-* ❓ str_replace_editor
-* ❓ tavily-tavily_search
-* ❓ str_replace_editor
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && grep -n "mcp-servers\|mcpServers" pkg/parser...`
-* ❓ str_replace_editor
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && grep -A 50 "ExtractMCPConfigurations" pkg/pa...`
-* ❓ str_replace_editor
-* ❓ tavily-tavily_search
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && cat .github/workflows/scout.md | head -25`
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && find .github/workflows -name "*.md" -exec gr...`
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && find pkg/cli/workflows -name "*.md" -exec gr...`
-* ❓ tavily-tavily_search
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && cat .github/workflows/shared/tavily-mcp.md`
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && find .github/workflows -name "*.md" -type f ...`
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && grep -r "ProcessImports" pkg/ --include="*.g...`
-* ❓ tavily-tavily_search
-* ❓ `cd /home/runner/work/gh-aw/gh-aw && ls -la .github/instructions/`
-* ❓ str_replace_editor
-* ❓ `cat > /tmp/gh-aw/cache-memory/research-mcp-imports.md << 'EOF' # MCP Server Imports Re...`
-* ❓ safe_outputs-add_comment
-* ❓ `cat > /tmp/gh-aw/cache-memory/mcp-imports-summary.txt << 'EOF' MCP Server Imports Rese...`
-
-## 📊 Information
-
-**Turns:** 48
-
-
-## 🤖 Reasoning
-
-I'll help research MCP server imports in compilers, focusing on how to support importing mcp-servers entries and understanding best practices for this feature.
-
-❓ tavily-tavily_search: MCP server Model Context Protocol import configuration compiler
-
-❓ tavily-tavily_search: Model Context Protocol mcp-servers configuration schema format
-
-❓ tavily-tavily_search: compiler import external configuration YAML frontmatter include directive
-
-❓ str_replace_editor: view
-
-Let me explore the repository structure and current implementation to better understand the context:
-
-❓ str_replace_editor: view
-
-❓ Find workflow markdown files: `find /home/runner/work/gh-aw/gh-aw -name "*.md" -path "*/workflows/*" | head -20`
-
-❓ Search for tavily references in workflows: `grep -r "tavily" /home/runner/work/gh-aw/gh-aw/.github/workflows/ --include="*.m...`
-
-❓ str_replace_editor: view
-
-❓ str_replace_editor: view
-
-❓ str_replace_editor: view
-
-❓ Find compiler and parser files: `find /home/runner/work/gh-aw/gh-aw/pkg -name "*.go" | grep -E "(parser|workflow|...`
-
-❓ tavily-tavily_search: GitHub Actions workflow YAML include import external configuration best practices
-
-❓ Find Go files with include/import logic: `find /home/runner/work/gh-aw/gh-aw/pkg -type f -name "*.go" | xargs grep -l "inc...`
-
-❓ str_replace_editor: view
-
-❓ tavily-tavily_search: MCP server configuration inheritance merge composition pattern
-
-❓ tavily-tavily_search: "mcp-servers" "tools" configuration YAML merge strategy deepmerge
-
-❓ Find parser files with import handling: `cd /home/runner/work/gh-aw/gh-aw && find pkg/parser -name "*.go" -exec grep -l "...`
-
-❓ str_replace_editor: view
-
-❓ Search for imports handling in frontmatter parser: `cd /home/runner/work/gh-aw/gh-aw && grep -n "imports" pkg/parser/frontmatter.go ...`
-
-❓ Look for ProcessImports function: `cd /home/runner/work/gh-aw/gh-aw && grep -A 20 "ProcessImports" pkg/parser/front...`
-
-❓ str_replace_editor: view
-
-❓ tavily-tavily_search: YAML deep merge strategy tools configuration inheritance frontmatter
-
-❓ str_replace_editor: view
-
-❓ Search for mcp-servers references: `cd /home/runner/work/gh-aw/gh-aw && grep -n "mcp-servers\|mcpServers" pkg/parser...`
-
-❓ str_replace_editor: view
-
-❓ Look for mcp-servers and tools handling in MCP extraction: `cd /home/runner/work/gh-aw/gh-aw && grep -A 50 "ExtractMCPConfigurations" pkg/pa...`
-
-❓ str_replace_editor: view
-
-❓ tavily-tavily_search: compiler configuration import reusable modules composition pattern best practices
-
-Let me now look at example workflow files and understand the current structure better:
-
-❓ View scout workflow example: `cd /home/runner/work/gh-aw/gh-aw && cat .github/workflows/scout.md | head -25`
-
-❓ Find workflows using mcp-servers: `cd /home/runner/work/gh-aw/gh-aw && find .github/workflows -name "*.md" -exec gr...`
-
-❓ Find test workflows using mcp-servers: `cd /home/runner/work/gh-aw/gh-aw && find pkg/cli/workflows -name "*.md" -exec gr...`
-
-❓ tavily-tavily_search: "import directive" "shared configuration" markdown frontmatter YAML merge deep
-
-Now let me check the documentation to understand the issue better:
-
-❓ View the tavily MCP include file: `cd /home/runner/work/gh-aw/gh-aw && cat .github/workflows/shared/tavily-mcp.md`
-
-❓ Find workflows with tools section: `cd /home/runner/work/gh-aw/gh-aw && find .github/workflows -name "*.md" -type f ...`
-
-❓ Find usages of ProcessImports function: `cd /home/runner/work/gh-aw/gh-aw && grep -r "ProcessImports" pkg/ --include="*.g...`
-
-Perfect! Now I have a good understanding. Let me search for more specific information about merge strategies and then compile the research report:
-
-❓ tavily-tavily_search: GitHub Actions reusable workflow configuration merge strategy composition
-
-❓ Check instructions directory: `cd /home/runner/work/gh-aw/gh-aw && ls -la .github/instructions/`
-
-❓ str_replace_editor: view
-
-Now let me write the memory cache and prepare the comprehensive research report:
-
-❓ Write research summary to cache memory: `cat > /tmp/gh-aw/cache-memory/research-mcp-imports.md << 'EOF' # MCP Server Imports Re...`
-
-❓ safe_outputs-add_comment: # 🔍 Scout Research Report
-
-*Triggered by @pelikhan*
-
-## Executive Summary
-
-The request is to move T...
-
-Perfect! The research report has been successfully posted. Let me also create a summary in the cache memory for future reference:
-
-❓ Create summary file in cache memory: `cat > /tmp/gh-aw/cache-memory/mcp-imports-summary.txt << 'EOF' MCP Server Imports Rese...`
-
diff --git a/actions/setup/js/test-fWIxhd/safe-outputs.jsonl b/actions/setup/js/test-fWIxhd/safe-outputs.jsonl
deleted file mode 100644
index 003597e20d..0000000000
--- a/actions/setup/js/test-fWIxhd/safe-outputs.jsonl
+++ /dev/null
@@ -1 +0,0 @@
-{"type":"create_issue","title":"Test Issue","body":"Test body"}
\ No newline at end of file
diff --git a/actions/setup/js/test-fWIxhd/test.log b/actions/setup/js/test-fWIxhd/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-fWIxhd/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-i7RCld/test.log b/actions/setup/js/test-i7RCld/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-i7RCld/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-lq2w64/test.log b/actions/setup/js/test-lq2w64/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-lq2w64/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file
diff --git a/actions/setup/js/test-noI9xR/test.log b/actions/setup/js/test-noI9xR/test.log
deleted file mode 100644
index 6b584e8ece..0000000000
--- a/actions/setup/js/test-noI9xR/test.log
+++ /dev/null
@@ -1 +0,0 @@
-content
\ No newline at end of file