perf: rest animations, cheaper chat typing and server polls, async per-tool hooks

UI
- No animation runs forever: Sagi moods, the custom pet sprite, status dots
  (pulse-dot, pulse-slow, ping) and the chat shimmer play a few cycles and
  rest. An infinite animation kept the browser drawing 60 frames a second.
- useSagiBrain: frames that only move lastActivityAt by <1 s keep the old
  state; the 1 s tick re-renders only when the mood changes.
- ConversationView and MessageList are memoized and PaneChat passes stable
  callbacks: typing in the chat box no longer re-renders the conversation.
- The chat's 700 ms menu poll reads plain text first and scans the terminal
  cell by cell (reusing one cell) only when a row could be a menu cursor row.

Server
- GET /sessions/:id/transcript?after=N skips lines <= N before JSON.parse.
- GET /sessions/:id/outline is cached until the session's events change.
- broadcastLane uses the cached tmux.liveSessions(); tmux.hasSession answers
  "yes" from the fresh shared list instead of a synchronous spawn (50-150 ms
  per psmux call on Windows, blocking terminal echo).
- pane-mode polls every 150 ms; writeHourly writes only changed hour buckets.
- /api/metrics: agenthub_event_loop_delay_seconds{quantile}.

Hooks
- PreToolUse/PostToolUse run with "async": true (settings installer,
  plugin.json, S3 managed settings): Claude Code no longer waits a Node start
  before and after every tool call.

Committed with --no-verify at the owner's request: the pre-commit full server
suite fails on 9 tests that fail identically on 0c2f498 (lane-sync merge: no
git identity on this host; read-pool; resolveHost; codex-usage 1 ms flake).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

server/lib/hourly-usage.js

@@ -11,11 +11,25 @@ const { db, stmts } = require("../db");
 const TranscriptCache = require("./transcript-cache");
 const { longArgs } = require("./token-usage");
 
-/** Write every hour of one transcript parse (MAX per hour, see the statement). */
+// What this process last wrote per session and hour bucket: a long session has
+// one bucket per active hour per model, and every hook re-sent all of them.
+// The upsert keeps the MAX, so skipping a bucket whose numbers did not move
+// writes exactly the same rows. Bounded (oldest session dropped first); a miss
+// only means one full write, never a lost one.
+const WRITTEN_MAX = 500;
+const written = new Map();
+
+/** Write the hours of one transcript parse that changed (MAX per hour, see the statement). */
 function writeHourly(sessionId, tokensByHour) {
-  // Runs every hour of the session on each hook — ~1 row per active hour per
-  // model; the statement leaves a row that did not grow untouched (no write).
-  for (const t of Object.values(tokensByHour || {})) {
+  let seen = written.get(sessionId);
+  if (seen) written.delete(sessionId);
+  else seen = new Map();
+  written.set(sessionId, seen);
+  if (written.size > WRITTEN_MAX) written.delete(written.keys().next().value);
+  for (const [key, t] of Object.entries(tokensByHour || {})) {
+    const sig = JSON.stringify(t);
+    if (seen.get(key) === sig) continue;
+    seen.set(key, sig);
     stmts.upsertHourlyUsage.run(
       sessionId,
       t.hour,
@@ -73,4 +87,11 @@ function backfillHourly() {
   });
 }
 
-module.exports = { writeHourly, backfillHourly };
+/** Forget what was written — for `ids`, or every session. Whatever deletes
+ *  hourly rows calls it, or the next hooks would skip re-writing them. */
+function forgetHourly(ids = null) {
+  if (!ids) return written.clear();
+  for (const id of ids) written.delete(id);
+}
+
+module.exports = { writeHourly, backfillHourly, forgetHourly };