feat(llm): add LiteLLM provider — enum, client, tests, QA follow-ups

- Add LITELLM = "litellm" to LLMProviderType enum (already in HEAD from prev commit)
- Add comment in LLMClient.__init__ explaining standard Bearer auth for LiteLLM
- Add regression test test_provider_type_enum_values — guard against value drift
- Add test_litellm_client_uses_default_bearer_auth — verify no extra headers
- Clean up duplicate @RELATION lines in models.py
- All 20 backend + 4 frontend tests pass

backend/src/plugins/llm_analysis/__tests__/test_client_headers.py

@@ -28,4 +28,28 @@ def test_openrouter_client_includes_referer_and_title_headers(monkeypatch):
     assert headers["HTTP-Referer"] == "http://localhost:8000"
     assert headers["X-Title"] == "ss-tools-test"
 # endregion test_openrouter_client_includes_referer_and_title_headers
+
+
+# region test_litellm_client_uses_default_bearer_auth [TYPE Function]
+# @RELATION: BINDS_TO -> TestClientHeaders
+# @PURPOSE: LiteLLM proxy uses standard OpenAI-compatible Bearer auth — no special headers needed.
+# @PRE: Client is initialized for LITELLM provider.
+# @POST: Async client headers include only Authorization — no extra provider-specific headers.
+def test_litellm_client_uses_default_bearer_auth():
+    """Verify LiteLLM client initialization uses standard Bearer auth without extra headers."""
+    client = LLMClient(
+        provider_type=LLMProviderType.LITELLM,
+        api_key="sk-litellm-key-123456",
+        base_url="http://localhost:4000/v1",
+        default_model="gpt-4o",
+    )
+
+    headers = dict(client.client.default_headers)
+    assert headers["Authorization"] == "Bearer sk-litellm-key-123456"
+    # LiteLLM proxy uses standard OpenAI-compatible auth — no special headers
+    assert "HTTP-Referer" not in headers
+    assert "X-Title" not in headers
+    assert "Authentication" not in headers
+    assert "X-API-Key" not in headers
+# endregion test_litellm_client_uses_default_bearer_auth
 # endregion TestClientHeaders

backend/src/plugins/llm_analysis/models.py

@@ -2,8 +2,6 @@
 # @BRIEF Define Pydantic models for LLM Analysis plugin.
 # @LAYER: Domain
 # @RELATION DEPENDS_ON -> pydantic
-# @RELATION DEPENDS_on -> pydantic
-# @RELATION DEPENDs_on -> pydantic
 
 from datetime import datetime
 from enum import Enum

backend/src/plugins/llm_analysis/service.py

@@ -711,6 +711,9 @@ class LLMClient:
         if self.provider_type == LLMProviderType.KILO:
             default_headers["Authentication"] = f"Bearer {self.api_key}"
             default_headers["X-API-Key"] = self.api_key
+        # LiteLLM proxy uses standard OpenAI-compatible Bearer auth — no special headers needed.
+        # It routes to upstream providers transparently, and the default Authorization header
+        # is sufficient. No additional headers like HTTP-Referer or X-API-Key are required.
 
         http_client = httpx.AsyncClient(headers=default_headers, timeout=120.0)
         self.client = AsyncOpenAI(