← Back to Skills Marketplace
Token Budget
by
OpenAuthority
· GitHub ↗
· v1.0.2
· MIT-0
147
Downloads
3
Stars
0
Active Installs
3
Versions
Install in OpenClaw
/install token-budget
Description
Track your AI agent's token usage, API spend, and set soft budget thresholds with in-session warnings.
Usage Guidance
This skill appears to do what it says: estimate token usage from the OpenClaw session and store soft thresholds/history in a local file. Before installing, check the following: 1) Inspect or create ~/.openclaw/openauthority/budget-state.json and confirm it contains only non-sensitive settings (thresholds, history), not API keys or secrets. 2) Because the registry metadata did not declare the config path the SKILL.md uses, ask the publisher to update the metadata to list that path explicitly so you know what will be read/written. 3) Remember SKILL.md claims no external data exfiltration, but instruction-only skills cannot enforce behavior — if you need guaranteed enforcement, use the platform-level OpenAuthority plugin or policy controls for hard budget limits. 4) If you want higher assurance, request a code-backed implementation or source repo (so you can review file I/O behavior) or a sample budget-state.json schema. If those checks are acceptable, the skill's accesses are proportionate and coherent with its purpose.
Capability Analysis
Type: OpenClaw Skill
Name: token-budget
Version: 1.0.2
The 'token-budget' skill is a utility designed to track AI token usage and estimated API costs. It operates by reading session metadata and maintaining a local state file (~/.openclaw/openauthority/budget-state.json). The instructions in SKILL.md are consistent with the stated purpose, and there is no evidence of data exfiltration, malicious execution, or harmful prompt injection.
Capability Assessment
Purpose & Capability
Name, description, and runtime instructions all align: the skill only needs session token metadata and a small local state file to store thresholds/history. There are no unexpected credentials or unrelated capabilities requested.
Instruction Scope
SKILL.md limits actions to reading OpenClaw session metadata and a local state file (~/.openclaw/openauthority/budget-state.json), estimating costs, and printing warnings. That scope is appropriate for the described purpose. Note: the skill claims 'No data is sent externally' — that is an instruction-level assertion (there is no code to verify), so you should trust the runtime environment to enforce that. Also the SKILL.md references the local config path while the registry metadata did not declare any required config paths (inconsistency).
Install Mechanism
Instruction-only skill with no install spec and no code files to execute. This minimizes installation risk because nothing will be downloaded or installed by the skill itself.
Credentials
The skill requests no environment variables or credentials. The only persistent access is to a single local state file under the user's home directory to store thresholds and history — this is proportionate to its purpose.
Persistence & Privilege
The skill is not always-on and does not request elevated platform privileges. It does rely on a local state file for persistence, which is reasonable for storing user-configured thresholds. The metadata did not declare this config path, which would be useful to see declared explicitly.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install token-budget - After installation, invoke the skill by name or use
/token-budget - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.2
token-budget 1.0.2 is a maintenance release with no code or documentation changes.
- Internal version bump from 1.0.1 to 1.0.2
- No file changes detected; functionality and documentation remain the same
v1.0.1
- Alerting now prints warnings directly in the session when a threshold is crossed, instead of supporting external channels like Telegram.
- Updated skill description and documentation to remove references to multi-channel alerting.
- Clarified that all tracking is local and read-only from session context.
- No code changes; documentation only.
v1.0.0
Initial release of the token-budget skill.
- Track AI agent's token usage and estimated API spend in real time
- Set soft daily token thresholds and receive usage alerts via Telegram or console
- View detailed session summaries and cost breakdowns by model and time window
- Review daily spend and token history for the past 7 days
- All tracking is local; no external data sharing
Metadata
Frequently Asked Questions
What is Token Budget?
Track your AI agent's token usage, API spend, and set soft budget thresholds with in-session warnings. It is an AI Agent Skill for Claude Code / OpenClaw, with 147 downloads so far.
How do I install Token Budget?
Run "/install token-budget" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Token Budget free?
Yes, Token Budget is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Token Budget support?
Token Budget is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Token Budget?
It is built and maintained by OpenAuthority (@openauthority); the current version is v1.0.2.
More Skills