fix(si): drop fabricated per-model weekly-remaining; tighten statusline line 5

AGENTS.md

@@ -1,7 +1,7 @@
 <!-- gitnexus:start -->
 # GitNexus MCP
 
-This project is indexed by GitNexus as **claude-session-intelligence** (698 symbols, 2165 relationships, 57 execution flows).
+This project is indexed by GitNexus as **claude-session-intelligence** (704 symbols, 2183 relationships, 58 execution flows).
 
 ## Always Start Here
 

CLAUDE.md

@@ -71,7 +71,7 @@ Disable with `toolArchive.enabled=false` or `CLAUDE_TOOL_ARCHIVE=0` if you don't
 <!-- gitnexus:start -->
 # GitNexus MCP
 
-This project is indexed by GitNexus as **claude-session-intelligence** (698 symbols, 2165 relationships, 57 execution flows).
+This project is indexed by GitNexus as **claude-session-intelligence** (704 symbols, 2183 relationships, 58 execution flows).
 
 ## Always Start Here
 

README.md

@@ -293,7 +293,7 @@ Set `fields` in `~/.claude/statusline-intel.json` to the list + order you want.
 | `compactCost` | `c$12.34 / s$68.20` | Cost + cache-savings accrued **since last /compact**. Incrementally cached at `/tmp/claude-cost-sincecompact-<sid>`, auto-invalidated when compact mtime advances. Shows `c$` only when ≥$0.01, `s$` only when ≥$0.10; hidden when both trivial (dim) |
 | `blockUsage` | `b:47% r:1hr 12m` | Claude's 5-hour block quota used + reset countdown; leads line 2 next to session duration + time until reset. Sourced from statusline stdin `rate_limits` (CC ≥2.1.90); cached `/api/oauth/usage` fallback for older builds, rendering last-good values with a `~` marker during API outages |
 | `weekUsage` | `w:31% r:4d 12hr` | Weekly (7-day) quota used + time until reset. Same sourcing as `blockUsage`; % zone-coloured: green <60%, yellow 60-85%, orange 85-95%, red ≥95% |
-| `modelSplit` | `fable-5:sc$80·sp99%·wr:~85%` | Per model: session cost (`sc$`), share of session spend (`sp%`), and estimated weekly window remaining (`wr:~%`). `wr` attributes the account weekly used% across models by their share of this week's tracked spend — an SI estimate (hence `~`), since Anthropic exposes no per-model weekly quota. Parent-session splits are recorded to `session_model_costs` on each redraw, so the weekly attribution gets richer as sessions run |
+| `modelSplit` | `fable-5:$1.7k·61% opus-4-8:$938·33%` | Per model: session spend (`$`, abbreviated to `k` past $1000) and share of session spend (`%`). Merges parent-conversation turns with subagent runs from the events DB; top 4 models, ones rounding to 0% dropped. No per-model quota is shown — Anthropic exposes only the account-wide weekly window (the `w:` cell), so a per-model "weekly remaining" would be fabricated. Per-model session spend is still recorded to `session_model_costs` for `/si stats` |
 | `tokenSpeed` | `tps:238` | Output-token generation speed (tokens/sec) — median of the last few streamed turns, measured from per-message chunk-timestamp spans (pure generation time). Falls back to session average (total output / total API time) when the tail has no streamed turn |
 | `modelQuota` | `opus:12/88%` | Per-model QUOTA used/available — dormant until Anthropic populates per-model limits (stdin extra `rate_limits` keys or `seven_day_<model>` buckets in the usage payload; both are null on plans without per-model limits) |
 | `cwd` | `~/DWS/claude-session-intelligence` | Full working directory, ccstatusline-style. `$HOME` collapses to `~`; middle-ellipsis when longer than 60 chars (keeps the leaf) (dim) |

plugins/session-intelligence/statusline/statusline-intel.js

@@ -712,6 +712,17 @@ function fmtUsd(n) {
   return `$${n.toFixed(2)}`;
 }
 
+// Denser USD for the per-model split, where a long session stacks several
+// figures on one line: roll to `k` past $1000 so `$1727` reads `$1.7k`,
+// `$12480` reads `$12k`. Small values keep cents.
+function fmtUsdShort(n) {
+  if (!Number.isFinite(n) || n <= 0) return '$0';
+  if (n >= 10000) return `$${Math.round(n / 1000)}k`;
+  if (n >= 1000)  return `$${(n / 1000).toFixed(1)}k`;
+  if (n >= 10)    return `$${Math.round(n)}`;
+  return `$${n.toFixed(2)}`;
+}
+
 // Single-letter duration formatter for tight cells (compactAge, session).
 // `3hr 23m` -> `3h23m`, `11hr` -> `11h`, `45m` -> `45m`. Spans over 1d
 // keep the day component and skip minutes — the precision isn't useful at
@@ -1519,7 +1530,7 @@ function buildRenderers(C) {
       return `${C.dim}tps:${v}${C.reset}`;
     },
 
-    modelSplit: (input, ctx) => {
+    modelSplit: (_input, ctx) => {
       // Parent-conversation turns from the transcript accumulator...
       const byModel = (ctx.tokenTotals && ctx.tokenTotals.byModel) || {};
       const merged = { ...byModel };
@@ -1534,45 +1545,28 @@ function buildRenderers(C) {
       if (!entries.length || total <= 0) return '';
       entries.sort((a, b) => b[1] - a[1]);
 
-      const events = loadEventsLib();
-      // Persist this session's per-model split so weeklyModelSpend can see
-      // parent-session spend across sessions (agents are already in the DB).
+      // Persist this session's per-model split for cross-session analytics
+      // (/si stats) — the only surviving record of parent-session splits
+      // once a transcript ages out.
       try {
+        const events = loadEventsLib();
         if (events && typeof events.upsertSessionModelCosts === 'function' && ctx.sessionId) {
           events.upsertSessionModelCosts(ctx.sessionId, byModel);
         }
       } catch { /* telemetry is best-effort */ }
 
-      // Per-model weekly-remaining ESTIMATE (`wr:~61%`): Anthropic exposes
-      // no per-model weekly quota (seven_day_<model> buckets are null on
-      // plans without per-model limits — verified 2026-07-17), so attribute
-      // the account weekly used% across models by their share of this
-      // week's tracked spend: wr(model) = 100 − weeklyUsed × spendShare.
-      // `~` marks it as SI's estimate, not an Anthropic number.
-      let weeklyUsed = null;
-      const sd = input && input.rate_limits && input.rate_limits.seven_day;
-      if (sd && typeof sd.used_percentage === 'number') weeklyUsed = sd.used_percentage;
-      let weekSpend = null;
-      if (weeklyUsed !== null && events && typeof events.weeklyModelSpend === 'function') {
-        try { weekSpend = events.weeklyModelSpend(); } catch { weekSpend = null; }
-      }
-      const weekTotal = weekSpend ? weekSpend.reduce((sum, r) => sum + (r.cost_usd || 0), 0) : 0;
-      const weekByModel = weekSpend
-        ? Object.fromEntries(weekSpend.map((r) => [r.model, r.cost_usd || 0]))
-        : {};
-
-      const parts = entries.slice(0, 4).map(([m, c]) => {
+      // Self-labelling `$·%`: session spend + share of session spend. No
+      // per-model quota shown — Anthropic exposes only the account-wide
+      // weekly window (the separate `w:` cell), so a per-model "weekly
+      // remaining" would be fabricated. Models rounding to 0% are dropped.
+      const parts = [];
+      for (const [m, c] of entries.slice(0, 4)) {
+        const share = Math.round((c / total) * 100);
+        if (share < 1) continue;
         const name = m.replace(/^claude-/, '').replace(/-20\d{6,}$/, '');
-        const usd = fmtUsd(c);
-        let cell = `${name}:sc${usd || '$0'}\u00b7sp${Math.round((c / total) * 100)}%`;
-        if (weeklyUsed !== null && weekTotal > 0) {
-          const share = (weekByModel[m] || 0) / weekTotal;
-          const wr = Math.max(0, Math.min(100, Math.round(100 - weeklyUsed * share)));
-          cell += `\u00b7wr:~${wr}%`;
-        }
-        return cell;
-      });
-      return `${C.dim}${parts.join(' ')}${C.reset}`;
+        parts.push(`${name}:${fmtUsdShort(c)}·${share}%`);
+      }
+      return parts.length ? `${C.dim}${parts.join(' ')}${C.reset}` : '';
     },
 
     /** Per-model QUOTA used/available — dormant until real data exists.