docs(env): add missing FAL_AI_API_KEY alias, REPLICATE_API_TOKEN, HIGGSFIELD credentials

The provider tools read these env vars but .env.example never listed
them, so a fresh setup silently misses them:

- FAL_AI_API_KEY — read by kling/minimax/veo/seedance/recraft/flux
  as an alias for FAL_KEY (os.environ.get('FAL_KEY') or FAL_AI_API_KEY),
  but only FAL_KEY was documented.
- REPLICATE_API_TOKEN — required by seedance_replicate; without it the
  Replicate-backed Seedance path stays unselectable.
- HIGGSFIELD_API_KEY / _API_SECRET / HIGGSFIELD_KEY — the three forms
  higgsfield_video.py accepts (combined key:secret vs split pair).

Doc-only, zero runtime change.

Refs: docs/REVIEW-image-to-video-voice.md §8 #11

Co-Authored-By: Claude <noreply@anthropic.com>

.env.example

@@ -4,6 +4,17 @@
 # --- Image + video gateway ---
 FAL_KEY=                     # FLUX images, Google Veo video, Kling video, MiniMax video, Recraft images
                              # Get one at https://fal.ai/dashboard/keys
+FAL_AI_API_KEY=              # Alias for FAL_KEY (some SDKs/docs use this name); either one is read.
+
+# --- Replicate ---
+REPLICATE_API_TOKEN=         # Replicate-hosted video gen (seedance_replicate). Needed to make the
+                             # Replicate-backed Seedance path selectable alongside the fal.ai one.
+                             # Get one at https://replicate.com/account/api-tokens
+
+# --- Higgsfield ---
+HIGGSFIELD_API_KEY=          # Higgsfield Cloud key (higgsfield_video). Pair with the secret below,
+HIGGSFIELD_API_SECRET=       # or use the combined HIGGSFIELD_KEY="<key>:<secret>" form instead.
+# HIGGSFIELD_KEY=            # Combined key:secret — set this INSTEAD of the _KEY/_SECRET pair if you prefer.
 
 # --- Google (one key unlocks image gen + TTS) ---
 GOOGLE_API_KEY=              # Google Imagen images, Google Cloud TTS (700+ voices, 50+ languages)