fix(meta-ai): declare supports_vision_tool_messages=False — Muse Spark 400s on image tool results

Muse Spark accepts images on user turns but returns HTTP 400
invalid_request_error 'messages[N].content did not match any supported
type' when the vision_analyze multimodal envelope lands in a role:tool
message. With the profile veto now honored by the vision fast-path gates,
declaring the limitation routes tool-result images through the aux-LLM
text path while user-message vision stays enabled.

Fixes #101668
Refs #47742

plugins/model-providers/meta-ai/__init__.py

@@ -121,6 +121,9 @@ meta_ai = MetaAIProfile(
     api_mode="codex_responses",
     # Muse Spark is natively multimodal (image/video/pdf/audio in, text out).
     supports_vision=True,
+    # ...but only on user turns: an image envelope inside a role:tool message
+    # 400s "messages[N].content did not match any supported type" (#101668).
+    supports_vision_tool_messages=False,
     # Cheap contributor tier is a good default for auxiliary tasks
     # (compaction, title generation, vision) when this is the main provider.
     default_aux_model="muse-spark-1.2-contributor",

tests/providers/test_meta_ai_profile.py

@@ -28,6 +28,8 @@ class TestMetaAIProfile:
         assert p.api_mode == "codex_responses"
         assert "MODEL_API_KEY" in p.env_vars
         assert p.supports_vision is True
+        # Images are accepted on user turns only; tool-result envelopes 400 (#101668).
+        assert p.supports_vision_tool_messages is False
         assert p.default_aux_model == "muse-spark-1.2-contributor"
         assert p.default_max_tokens == 16384
         assert p.fallback_models == ("muse-spark-1.2",)