fix(model-resolver): use first fallback entry when model cache unavailable

When availableModels is empty (no cache in CI), use the first entry
from fallbackChain directly instead of falling back to systemDefault.
This ensures categories and agents use their configured models even
when the model cache file doesn't exist.

Fixes:
- model-resolution check returning 'warn' instead of 'pass' in CI
- DEFAULT_CATEGORIES not being used when no cache available
- Unstable agent detection failing (models falling back to non-gemini)

src/agents/utils.test.ts

@@ -45,17 +45,17 @@ describe("createBuiltinAgents with model overrides", () => {
     expect(agents.Sisyphus.reasoningEffort).toBeUndefined()
   })
 
-  test("Oracle uses system default when no availableModels provided", async () => {
-    // #given - no available models, falls back to system default
+  test("Oracle uses first fallback entry when no availableModels provided (no cache scenario)", async () => {
+    // #given - no available models simulates CI without model cache
 
     // #when
     const agents = await createBuiltinAgents([], {}, undefined, TEST_DEFAULT_MODEL)
 
-    // #then - falls back to system default (anthropic/claude-opus-4-5)
-    expect(agents.oracle.model).toBe("anthropic/claude-opus-4-5")
-    expect(agents.oracle.thinking).toEqual({ type: "enabled", budgetTokens: 32000 })
-    expect(agents.oracle.reasoningEffort).toBeUndefined()
-    expect(agents.oracle.textVerbosity).toBeUndefined()
+    // #then - uses first fallback entry (openai/gpt-5.2) instead of system default
+    expect(agents.oracle.model).toBe("openai/gpt-5.2")
+    expect(agents.oracle.reasoningEffort).toBe("medium")
+    expect(agents.oracle.textVerbosity).toBe("high")
+    expect(agents.oracle.thinking).toBeUndefined()
   })
 
   test("Oracle with GPT model override has reasoningEffort, no thinking", async () => {