adventure: author Pete's dispatches in his voice with the local LLM

emitFact now runs the final fact through authorDispatch, which asks the local
Ollama model for a warm-reporter headline and lede and ships them on the Fact.
Pete guards and publishes them, falling back to its own templates on anything
it rejects — so authoring is best-effort by design: LLM off, a timeout, a
malformed generation, or an over-length pair all return an empty prose pair and
Pete templates the fact. The names allowed in the prose are the fact's Actors,
built from the post-opt-out subject/opponent, so what the model may say and what
Pete's guard permits are the same list. Synchronous like the holdem tip rewrite,
but on a tight 15s budget: news facts are infrequent and a template now beats a
voiced dispatch late. With no route for Pete to call back into this box, the
voice lives in the prompt here rather than in a Pete-owned inference endpoint.

internal/peteclient/client.go

@@ -48,6 +48,13 @@ type Fact struct {
 	Milestone  string   `json:"milestone,omitempty"`
 	OccurredAt int64    `json:"occurred_at"`
 	NoPush     bool     `json:"no_push,omitempty"` // backfill: suppress Pete web-push
+	// Headline/Lede are LLM-authored prose for this fact, both optional. Pete
+	// prefers them over its own template when present and past its prose-guard,
+	// and falls back to the template otherwise — so an empty pair (LLM off, or
+	// authoring failed) is the normal, safe case. Populated by emitFact; see
+	// authorDispatch. Names in the prose must come only from Actors.
+	Headline string `json:"headline,omitempty"`
+	Lede     string `json:"lede,omitempty"`
 }
 
 // Config controls the seam. Enabled=false makes Emit a durable no-op (nothing