fix: give remove-background job timeouts an actionable failure message (#518)

Extends the worker.ts timeout failure detail with actionable guidance (first-run model download, input too large for CPU inference, or busy/unavailable worker) while preserving the "Timed out after Ns" prefix so error_code classification, SSE terminal replay, and existing timeout assertions keep working. Adds a test assertion for the guidance.

Fixes #494

apps/api/src/jobs/worker.ts

@@ -369,7 +369,7 @@ async function processToolJob(job: Job<ToolJobData>): Promise<ToolJobResult> {
       const finalError = isCanceled
         ? "Canceled"
         : isTimeout
-          ? `Timed out after ${Math.round(timeoutMs / 1000)}s`
+          ? `Timed out after ${Math.round(timeoutMs / 1000)}s. On the first run the background-removal model may still be downloading; the image may be too large for CPU inference; or the worker may be busy or unavailable. Retry once the model has downloaded, or try a smaller image.`
           : errorMessage;
 
       // Log genuine processing faults at error level (clients only ever see

tests/integration/platform/worker-timeout.test.ts

@@ -116,6 +116,10 @@ describe("Worker timeout classification", () => {
     // Error message must mention timeout
     const error = finalRow?.error as { message: string };
     expect(error.message).toMatch(/timed out after 1s/i);
+    // Fail-fast message must include actionable guidance (issue #494)
+    expect(error.message).toMatch(
+      /model may still be downloading|too large for CPU|worker may be busy/i,
+    );
 
     // Both attempts ran (attempts column is set at the start of each attempt)
     expect(finalRow?.attempts).toBe(2);