feat(coding-agent): improved Python startup performance with prelude caching and shared gateway mode

- Added Python prelude caching to improve startup performance by caching helper functions and module discovery results.
- Added OMP_DEBUG_STARTUP environment variable for detailed startup stage tracing across initialization pipeline.
- Changed Python kernel initialization to require shared gateway mode, removing support for local kernel gateway startup.
- Improved error handling in shared gateway acquisition with automatic retries on 5xx status codes.
- Added repro-stuck.ts script for reproducing and debugging rare CLI hang issues with Chrome DevTools integration.
- Added trace-loader.ts preload script for tracing module resolution during startup diagnostics.

scripts/trace-loader.ts

@@ -0,0 +1,33 @@
+/**
+ * Bun preload script that traces module resolution.
+ * Usage: bun --preload ./scripts/trace-loader.ts <script>
+ */
+
+const startTime = performance.now();
+const resolved = new Set<string>();
+
+Bun.plugin({
+	name: "trace-loader",
+	setup(build) {
+		// Trace module resolution (doesn't interfere with loading)
+		build.onResolve({ filter: /.*/ }, (args) => {
+			// Skip if already traced this path
+			if (resolved.has(args.path)) {
+				return undefined;
+			}
+			resolved.add(args.path);
+			
+			const elapsed = (performance.now() - startTime).toFixed(1);
+			// Only trace local/project files, not node_modules
+			if (!args.path.includes("node_modules") && !args.path.startsWith("node:")) {
+				const shortPath = args.path.replace(process.cwd(), ".");
+				process.stderr.write(`[${elapsed}ms] resolve: ${shortPath}\n`);
+			}
+			
+			// Return undefined to let Bun handle resolution normally
+			return undefined;
+		});
+	},
+});
+
+process.stderr.write(`[trace-loader] preload active\n`);