claude-code-cache-fix 1.7.1 → 1.7.2

package/README.md

@@ -105,6 +105,42 @@ The module intercepts `globalThis.fetch` before Claude Code makes API calls to `
 
 All fixes are idempotent — if nothing needs fixing, the request passes through unmodified. The interceptor is read-only with respect to your conversation; it only normalizes the request structure before it hits the API.
 
+## Status line — quota warnings in real time
+
+The interceptor writes quota state to `~/.claude/quota-status.json` on every API call. The included `tools/quota-statusline.sh` script reads this file and displays a live status line in Claude Code showing:
+
+- **Q5h %** with burn rate (%/min)
+- **Q7d %** with burn rate (%/hr)
+- **TTL tier** — shows `TTL:1h` when healthy, **`TTL:5m` in red when the server has downgraded you** (typically at Q5h ≥ 100%)
+- **PEAK** in yellow during weekday peak hours (13:00–19:00 UTC)
+- **Cache hit rate %**
+- **OVERAGE** flag when active
+
+### Setup
+
+Copy the script and configure Claude Code to use it:
+
+```bash
+# Copy from the npm package to Claude Code's hooks directory
+mkdir -p ~/.claude/hooks
+cp "$(npm root -g)/claude-code-cache-fix/tools/quota-statusline.sh" ~/.claude/hooks/
+chmod +x ~/.claude/hooks/quota-statusline.sh
+```
+
+Add to `~/.claude/settings.json`:
+
+```json
+{
+  "statusLine": {
+    "command": "~/.claude/hooks/quota-statusline.sh"
+  }
+}
+```
+
+### Why this matters
+
+When the server downgrades your TTL to 5m (Layer 2 — quota-aware downgrade at Q5h ≥ 100%), **every idle longer than 5 minutes causes a full context rebuild**. Without the status line, this is invisible — you just notice things getting slower and more expensive. With the status line, the red `TTL:5m` warning tells you immediately: **stop working, wait for the Q5h window to reset, then resume**. Powering through overage compounds the drain; pausing breaks the cycle.
+
 ## Image stripping
 
 Images read via the Read tool are encoded as base64 and stored in `tool_result` blocks in conversation history. They ride along on **every subsequent API call** until compaction. A single 500KB image costs ~62,500 tokens per turn in carry-forward.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claude-code-cache-fix",
-  "version": "1.7.1",
+  "version": "1.7.2",
   "description": "Fixes prompt cache regression in Claude Code that causes up to 20x cost increase on resumed sessions",
   "type": "module",
   "exports": "./preload.mjs",

package/tools/quota-statusline.sh

@@ -0,0 +1,76 @@
+#!/bin/bash
+# Status line: show quota % and burn rate from claude-meter JSONL
+# Rate is calculated from window start (reset_time - window_size) to now
+# No prev file needed — each reading is self-contained
+
+input=$(cat)
+
+JSONL="$HOME/.claude/claude-meter.jsonl"
+
+if [ -f "$JSONL" ]; then
+  last=$(tail -1 "$JSONL" 2>/dev/null)
+
+  result=$(echo "$last" | python3 -c "
+import sys, json
+from datetime import datetime, timezone
+
+r = json.load(sys.stdin)
+q5h = int(r['q5h'] * 100)
+q7d = int(r.get('q7d', 0) * 100)
+overage = r.get('qoverage', '')
+ts = r.get('ts', '')
+q5h_reset = r.get('q5h_reset', 0)
+q7d_reset = r.get('q7d_reset', 0)
+
+now = datetime.fromisoformat(ts.replace('Z', '+00:00'))
+
+# Q5h: 5-hour window, rate = pct / minutes elapsed since window start
+rate5 = ''
+if q5h_reset > 0:
+    window_start = datetime.fromtimestamp(q5h_reset, tz=timezone.utc) - __import__('datetime').timedelta(hours=5)
+    elapsed_min = (now - window_start).total_seconds() / 60
+    if elapsed_min > 1 and q5h > 0:
+        rate5 = '{:+.1f}'.format(q5h / elapsed_min)
+
+# Q7d: 7-day window
+rate7 = ''
+if q7d_reset > 0:
+    window_start_7d = datetime.fromtimestamp(q7d_reset, tz=timezone.utc) - __import__('datetime').timedelta(days=7)
+    elapsed_min_7d = (now - window_start_7d).total_seconds() / 60
+    if elapsed_min_7d > 1 and q7d > 0:
+        rate7 = '{:+.1f}'.format(q7d / (elapsed_min_7d / 60))
+
+label = 'Q5h: {}%'.format(q5h)
+if rate5:
+    label += ' ({}%/m)'.format(rate5)
+label += ' | Q7d: {}%'.format(q7d)
+if rate7:
+    label += ' ({}%/hr)'.format(rate7)
+if overage == 'active':
+    label += ' | OVERAGE'
+
+# Add TTL tier from quota-status.json (written by interceptor)
+import os, pathlib
+qs_path = pathlib.Path.home() / '.claude' / 'quota-status.json'
+try:
+    qs = json.load(open(qs_path))
+    ttl = qs.get('cache', {}).get('ttl_tier', '')
+    hit = qs.get('cache', {}).get('hit_rate', '')
+    if ttl:
+        if ttl == '5m':
+            label += ' | \033[31mTTL:5m\033[0m'  # red
+        else:
+            label += ' | TTL:' + ttl
+    if hit and hit != 'N/A':
+        label += ' ' + hit + '%'
+    peak = qs.get('peak_hour', False)
+    if peak:
+        label += ' | \033[33mPEAK\033[0m'  # yellow
+except:
+    pass
+
+print(label)
+" 2>/dev/null)
+
+  [ -n "$result" ] && echo "$result"
+fi