evaluate: switch to Elo ratings with skillratings crate

Replace sort-based ranking with proper Elo system:
- Each agent TYPE has a persistent Elo rating (agent-elo.json)
- Each matchup: pick two random types, grab a recent action from
  each, LLM compares, update ratings
- Ratings persist across daily evaluations — natural recency bias
  from continuous updates against current opponents
- K=32 for fast adaptation to prompt changes

Usage: poc-memory agent evaluate --matchups 30 --model haiku

Co-Authored-By: Kent Overstreet <kent.overstreet@linux.dev>

poc-memory/src/main.rs

@@ -568,9 +568,9 @@ enum AgentCmd {
     /// Evaluate agent quality by LLM-sorted ranking
     #[command(name = "evaluate")]
     Evaluate {
-        /// Samples per agent type
-        #[arg(long, default_value_t = 5)]
-        samples: usize,
+        /// Number of pairwise matchups to run
+        #[arg(long, default_value_t = 30)]
+        matchups: usize,
         /// Model to use for comparison (haiku or sonnet)
         #[arg(long, default_value = "haiku")]
         model: String,
@@ -798,8 +798,8 @@ fn main() {
                 => cli::agent::cmd_fact_mine(&path, batch, dry_run, output.as_deref(), min_messages),
             AgentCmd::FactMineStore { path } => cli::agent::cmd_fact_mine_store(&path),
             AgentCmd::ReplayQueue { count } => cli::agent::cmd_replay_queue(count),
-            AgentCmd::Evaluate { samples, model, dry_run }
-                => cli::agent::cmd_evaluate_agents(samples, &model, dry_run),
+            AgentCmd::Evaluate { matchups, model, dry_run }
+                => cli::agent::cmd_evaluate_agents(matchups, &model, dry_run),
         },
 
         // Admin