feat(pi): emit model context and output limits when set

The pi adapter writes contextWindow and maxTokens to models.json only
when the profile sets them. Pi keeps its own defaults for every other
model. Tests cover both cases.

internal/adapters/pi/pi.go

@@ -116,6 +116,17 @@ func piModelsJSON(b *canonical.Bundle) ([]byte, error) {
 		if m.Alias != "" {
 			entry["name"] = m.Alias
 		}
+		// Emit limits only when the profile sets them explicitly. Pi applies
+		// its own defaults (contextWindow 128000, maxTokens 16384) to a model
+		// definition that omits them, so writing ContextLimit()/OutputLimit()
+		// unconditionally would push canonical's defaults onto every model
+		// that never asked for them.
+		if m.Context > 0 {
+			entry["contextWindow"] = m.Context
+		}
+		if m.Output > 0 {
+			entry["maxTokens"] = m.Output
+		}
 		modelEntries = append(modelEntries, entry)
 	}
 	if len(modelEntries) == 0 && b.Profile.Gateway.DefaultModel != "" {