Install
openclaw skills install @decisionvex/claude-usage-companionMonitor your Claude programmatic credit pool and align your interactive window — on your always-on box. Since June 2026 Anthropic meters non-interactive usage (Claude Agent SDK, `claude -p`, Claude Code GitHub Actions, third-party agents) from a SEPARATE monthly credit pool ($20 Pro / $100 Max 5x / $200 Max 20x, API rates, no rollover) instead of your 5-hour interactive window. claude-usage-companion (1) tracks that credit pool — month-to-date burn, projected month-end spend, and an alert before you run dry so automations don't stop unexpectedly; and (2) reminds you, the human, at your anchor time to send your first interactive prompt so your 5-hour window lines up with your working day. Use this skill when the user mentions Claude usage limits, the programmatic/Agent-SDK credit pool, running out of credits, monitoring claude -p or GitHub Actions spend, ccusage, usage reports, budgeting Claude automation on a VPS, or wants to align/start their interactive session window. It only reports and reminds — it never automates the model or circumvents any limit.
openclaw skills install @decisionvex/claude-usage-companionKeep your Claude usage on an always-on box healthy after the June 2026 usage split — when
programmatic usage (claude -p, Agent SDK, GitHub Actions, third-party agents) moved onto a
separate monthly credit pool instead of your interactive 5-hour window.
It does two clean, hands-off jobs:
Replaces
claude-session-warmer. That skill primed the interactive window withclaude -p, which no longer works: as of ~June 15 2026claude -pdraws from the separate credit pool and does not anchor the 5-hour interactive window. This companion is the honest successor — it never automates the model and never circumvents a limit; it only reports and reminds.
No automated priming of the window, no scripted interactive input, no token/credential tricks,
no limit circumvention. Usage data is read locally with ccusage (a parse of ~/.claude — no
model calls, no window side effects). Cost is USD at API rates, the same basis as the credit pool.
An always-on box (the VPS where you run OpenClaw / Claude Code) with Node 18+ and
ccusage. Fast path: npm i -g ccusage.
Run on the box, from the skill folder.
1. Tell it your timezone, plan, and start time.
cp config.example.json config.json
# edit: "timezone" (e.g. "Africa/Johannesburg"), "plan" (pro | max5x | max20x),
# and "anchor" (the local time you start work, e.g. "08:00")
2. Check the box and see your burn.
node bin/usage-companion.mjs check # verifies node + ccusage + reads your usage
node bin/usage-companion.mjs report # month-to-date burn, projection, top models
3. Schedule it.
node bin/usage-companion.mjs install # prints a ready-to-paste cron block
crontab -e # paste it, save
Now the box checks your credit pool a few times a day (alerting only when you cross a threshold) and reminds you at your anchor time to start your interactive session.
check — verify Node + ccusage are available and config is valid (run first).report — month-to-date spend, projected month-end, top models (no side effects).guard — evaluate thresholds; alert only if over warn_pct/critical_pct (what cron runs).remind — emit the anchor-time "start your session" nudge (what cron runs).status — show config and next scheduled actions.install [--cron-only] — print the cron block (--cron-only is pipeable to crontab -).config.json)| Field | Meaning | Default |
|---|---|---|
timezone | IANA tz for all clock math — set this | UTC |
plan | pro / max5x / max20x → sets the credit cap ($20/$100/$200) | max5x |
monthly_credit_usd | Override the cap if your number differs | null |
warn_pct / critical_pct | Alert thresholds (% of cap, actual or projected) | 80 / 95 |
anchor | Local time you start work (window reminder) | 08:00 |
reminder_enabled | Send the anchor-time nudge | true |
alert_command | Shell command to receive alerts on stdin (e.g. post to Slack). null = stdout + companion.log | null |
ccusage_cmd | How to invoke ccusage | ["npx","-y","ccusage@latest"] |
For your exact remaining balance, the authoritative source is the "Usage credits" line in
/usage inside Claude Code; report/guard approximate the pool from local ccusage cost, which
on an always-on automation box is essentially all programmatic.
bin/usage-companion.mjs — the engine (Node; only external tool is ccusage).config.example.json — copy to config.json and edit.README.md — public overview.