LaunchKit · 2026
Back to Skills

tokenguard

AI & LLMsv1.0.0

API cost guardian for AI agents.

3
1.6k downloads
by @g0head

Setup & Installation

openclaw skills install @g0head/tokenguard

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install tokenguard

What This Skill Does

TokenGuard tracks API spending per session, enforces configurable dollar limits, and blocks actions when budgets are exceeded. It logs every cost with timestamps and provides pre-flight budget checks before expensive calls.

Unlike ad-hoc manual tracking, TokenGuard enforces limits programmatically via exit codes, so scripts and agents can gate expensive operations without any custom logic.

When to use it

  • Stopping a runaway agent loop before it burns $200 overnight
  • Setting a $10 daily cap during development of a new AI feature
  • Auditing which operations consumed the most API budget last week
  • Scripting a CI pipeline that skips expensive calls when over budget
  • Extending a session limit mid-task without losing the spending history

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User asks: check if I can afford a $5 API call, then log the actual cost

AGENT
  1. 1Run `tokenguard status` to see current limit and remaining balance
  2. 2Run `tokenguard check 5.00` to verify the estimated cost fits the budget
  3. 3Call the external API if exit code is 0
  4. 4Run `tokenguard log 4.23 "Claude Sonnet - summarization"` to record the actual cost
  5. 5Run `tokenguard history` to confirm the entry was saved
OUTPUT

Spending log updated: $4.23 recorded, remaining balance displayed