feat: wire multi-agent config and routing

Co-authored-by: Mark Pors <1078320+pors@users.noreply.github.com>

docs/concepts/session-pruning.md

@@ -2,7 +2,7 @@
 summary: "Session pruning: tool-result trimming to reduce context bloat"
 read_when:
   - You want to reduce LLM context growth from tool outputs
-  - You are tuning agent.contextPruning
+  - You are tuning agents.defaults.contextPruning
 ---
 # Session Pruning
 
@@ -23,7 +23,7 @@ Session pruning trims **old tool results** from the in-memory context right befo
 Pruning uses an estimated context window (chars ≈ tokens × 4). The window size is resolved in this order:
 1) Model definition `contextWindow` (from the model registry).
 2) `models.providers.*.models[].contextWindow` override.
-3) `agent.contextTokens`.
+3) `agents.defaults.contextTokens`.
 4) Default `200000` tokens.
 
 ## Modes