feat(council): define litellm-backed tier policy

skills/council/references/roster.json

@@ -56,19 +56,25 @@
   },
   "tierPolicy": {
     "light": {
-      "advisorModel": null,
-      "synthesisModel": null,
-      "notes": "Use dedicated council agents as configured. No overrides by default."
+      "advisorModel": "litellm/gpt-5.3-codex",
+      "synthesisModel": "litellm/gpt-5.3-codex",
+      "advisorThinking": "low",
+      "synthesisThinking": "low",
+      "notes": "Use local LiteLLM for all council roles. Light tier = gpt-5.3-codex with low thinking."
     },
     "medium": {
-      "advisorModel": null,
-      "synthesisModel": null,
-      "notes": "Keep advisor agents cheap; rely on stronger dedicated synthesis agents."
+      "advisorModel": "litellm/gpt-5.3-codex",
+      "synthesisModel": "litellm/gpt-5.3-codex",
+      "advisorThinking": "high",
+      "synthesisThinking": "high",
+      "notes": "Use local LiteLLM for all council roles. Medium tier = gpt-5.3-codex with high thinking."
     },
     "heavy": {
-      "advisorModel": null,
-      "synthesisModel": "openai-codex/gpt-5.4",
-      "notes": "Preserve dedicated role identities. Escalate synthesis first; only override advisors when task risk justifies it."
+      "advisorModel": "litellm/gpt-5.4",
+      "synthesisModel": "litellm/gpt-5.4",
+      "advisorThinking": "high",
+      "synthesisThinking": "high",
+      "notes": "Use local LiteLLM for all council roles. Heavy tier = gpt-5.4 with high thinking."
     }
   }
 }