Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/athevon/tokeneater/agents-mdgit clone --depth 1 https://github.com/AThevon/TokenEaterWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/athevon/tokeneater/agents-md)<a href="https://agentmods.dev/instructions/athevon/tokeneater/agents-md"><img src="https://agentmods.dev/badge/instructions/athevon/tokeneater/agents-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.05722 | $0.05722 |
| Opus 5 | $0.02861 | $0.02861 |
| Sonnet 5 | $0.01144 | $0.01144 |
| Haiku 4.5 | $0.00572 | $0.00572 |
Grade C, and why
TokenEater AGENTS.md scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf ~/Library/Application\ Support/com.tokeneater.shared && \ How it starts
The opening of the file, as written. The whole thing — 250 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
TokenEater is a native macOS menu bar app, plus WidgetKit widgets and a floating overlay, that tracks Claude AI usage limits in real time. This file is the guide for contributors and AI coding agents: how the repo is laid out, how data flows, the rules that are easy to break, and the commands that build, test, and ship it.
It complements the other docs and deliberately does not duplicate them:
README.md- what the app does (full feature list), the security and privacy model, the public API shape.CONTRIBUTING.md- contribution workflow, commit conventions, how to get help, the external-contributor signing caveat.SETUP.md- building from source as an end user.docs/design/MASTER.mdanddocs/design/COLORING.md- the window design system and the Smart Color risk model.
Current version: 5.13.0 (MARKETING_VERSION in project.yml).
Language
- Everything on GitHub (issues, PRs, commits, branches) is in English.
- Conventional Commits are preferred (
feat:,fix:,chore:,docs:), since they make changelog generation easier. Don't stress about it; a wrong prefix can be fixed on merge.
Repo layout and targets
project.yml is the single source of truth for the project. The .xcodeproj is generated by XcodeGen and is gitignored. Never edit the .xcodeproj directly: those edits are lost on the next generate. To change targets, build settings, schemes, entitlements wiring, the installer prebuild step, or the version, edit project.yml then run xcodegen generate.
A stale
ClaudeUsageWidget.xcodeprojmay exist locally from before the rename. It is also generated and gitignored. Ignore it.
Three targets:
| Target | What it is | Sources |
|---|---|---|
TokenEaterApp |
The app host. Non-sandboxed (since v5.0) so it can read the token. LSUIElement (no Dock icon). |
TokenEaterApp/ + Shared/ |
TokenEaterWidgetExtension |
The WidgetKit extension. Sandboxed and read-only (WidgetKit requires app-sandbox: true). |
TokenEaterWidget/ + Shared/ |
TokenEaterTests |
Swift Testing bundle, unsigned. | TokenEaterTests/ + Shared/ + the single file TokenEaterApp/OnboardingViewModel.swift |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today Changed 438aa1f33e44
- 2d ago Changed 0a0691a5e32e
- 6d ago First seen · 250 lines · 5,722 tokens per session scan C cdeebfc91e6f
TokenEater AGENTS.md is an instructions file published in the GitHub repository AThevon/TokenEater (496 stars, last pushed today), licensed MIT. It adds 5,722 tokens to every session, about $0.0286 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
Applite CLAUDE.md
Instructions for milanvarady/Applite, covering claude.md, project overview, build & run, architecture and data flow.
OnlySwitch AGENTS.md
Instructions for jacklandrin/OnlySwitch, covering onlyswitch agent guidance, workflow, context routing and project-specific working rules.
Vocello AGENTS.md
AGENTS.md instructions for PowerBeef/Vocello, covering agents.md — vocello (qwenvoice), product and authority, start and resume work, release-first execution and hard invariants.
swift-tui AGENTS.md
AGENTS.md instructions for SwiftTUI/swift-tui, covering agents.md, build & test commands, architecture (one-page summary), development guidelines and code style.
skills-manager CLAUDE.md
Instructions for yibie/skills-manager, covering skills manager - development guide, tui engine: blessed, 核心原则, 参考文档 and 正确的 list 用法.
skills-manager AGENTS.md
Instructions for yibie/skills-manager, covering repository agent instructions and ui and interaction work.