fix(hyperframes-media): create the output dir for ElevenLabs TTS before writing [P1]

skills-manifest.json

@@ -42,7 +42,7 @@
       "files": 3
     },
     "hyperframes-media": {
-      "hash": "60f73ecb91858ea3",
+      "hash": "bd7512efa1a232d1",
       "files": 47
     },
     "hyperframes-registry": {

skills/hyperframes-media/scripts/lib/tts.mjs

@@ -216,6 +216,12 @@ export async function synthesizeOne({
 }) {
   if (provider === "heygen") return synthesizeHeygen({ text, voiceId, lang, speed, wavAbs });
   if (provider === "elevenlabs") {
+    // The Python helper writes straight to wavAbs; unlike heygen (transcodeToWav)
+    // and kokoro (the `hyperframes tts` CLI), it does NOT create the parent dir,
+    // so on a fresh project (no assets/voice/ yet) the save fails and the line is
+    // silently dropped as "TTS failed - omitted". Create it first, like the
+    // other providers do.
+    mkdirSync(dirname(wavAbs), { recursive: true });
     const { cmd, args } = pythonInvocation([
       "-c",
       ELEVENLABS_PY,

skills/hyperframes-media/scripts/lib/tts.test.mjs

@@ -1,9 +1,9 @@
 import { test } from "node:test";
 import assert from "node:assert/strict";
-import { mkdtempSync, writeFileSync, chmodSync, rmSync } from "node:fs";
-import { join } from "node:path";
+import { mkdtempSync, writeFileSync, chmodSync, rmSync, existsSync } from "node:fs";
+import { dirname, join } from "node:path";
 import { tmpdir } from "node:os";
-import { parseFfmpegDurationBanner, ffprobeDuration } from "./tts.mjs";
+import { parseFfmpegDurationBanner, ffprobeDuration, synthesizeOne } from "./tts.mjs";
 
 test("parseFfmpegDurationBanner reads ffmpeg's stderr Duration line", () => {
   const stderr = [
@@ -64,3 +64,25 @@ test("ffprobeDuration returns NaN when neither ffprobe nor ffmpeg resolve", () =
     rmSync(dir, { recursive: true, force: true });
   }
 });
+
+// Regression: the elevenlabs provider spawned Python to write straight to
+// wavAbs without creating its parent dir first (unlike heygen/kokoro), so a
+// fresh project with no assets/voice/ yet failed silently ("TTS failed -
+// omitted"). synthesizeOne must create the output dir before synthesizing —
+// asserted independently of whether the (keyless) Python actually produces a
+// file, since the mkdir runs before the spawn either way.
+test("synthesizeOne(elevenlabs) creates the output dir before writing", async () => {
+  const base = mkdtempSync(join(tmpdir(), "hf-tts-eleven-"));
+  const wavAbs = join(base, "assets", "voice", "line-0.wav"); // nested, not yet created
+  const originalKey = process.env.ELEVENLABS_API_KEY;
+  delete process.env.ELEVENLABS_API_KEY; // keeps the Python side a fast, clean failure
+  try {
+    // May resolve/spawn Python and fail (no key) — we only require the dir exists.
+    await synthesizeOne({ provider: "elevenlabs", text: "hi", voiceId: "v1", wavAbs });
+    assert.ok(existsSync(dirname(wavAbs)), "output directory should be created");
+  } finally {
+    if (originalKey === undefined) delete process.env.ELEVENLABS_API_KEY;
+    else process.env.ELEVENLABS_API_KEY = originalKey;
+    rmSync(base, { recursive: true, force: true });
+  }
+});