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/zghey/tokibean/agents-mdgit clone --depth 1 https://github.com/ZGhey/tokibeanWrote 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/zghey/tokibean/agents-md)<a href="https://agentmods.dev/instructions/zghey/tokibean/agents-md"><img src="https://agentmods.dev/badge/instructions/zghey/tokibean/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 | $0.03744 | $0.03744 |
| Opus 5 | $0.01872 | $0.01872 |
| Sonnet 5 | $0.00749 | $0.00749 |
| Haiku 4.5 | $0.00374 | $0.00374 |
Grade A, and why
tokibean AGENTS.md scanned grade A 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 5d ago.
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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- `official.rs` — fetches `https://api.anthropic.com/api/oauth/usage` via `curl` (token passed over curl's stdin with `-H @-`, never on the command line/process list). `get_token` uses **only the pet's own connected cred Copies of this mod
1 near-identical copy found in the catalogue:
- tokibean CLAUDE.md — 86% identical, 32 lines differ
How it starts
The opening of the file, as written. The whole thing — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
This file provides guidance to Codex (Codex.ai/code) when working with code in this repository.
Project
Tokibean 码豆 ("拱门·墩墩"): a Tauri 2 desktop pet that monitors Codex activity (via hooks) and token usage (via local JSONL logs, or the official Anthropic usage API). Transparent, always-on-top, frameless window with a tray icon. Cross-platform (Windows/macOS/Linux), though several features (WSL sync, Windows Credential Manager) are Windows-specific. Code comments and developer logs are in English. In-app UI text is bilingual: wrap user-facing strings with i18n::t("中文", "English") (backend, src-tauri/src/i18n.rs) or the t("key") dictionary (frontend, src/main.js) so the UI follows the system language — Chinese for zh* locales, English otherwise. Exception: the pet's tool-animation status keys (cmd/reading/coding/…, from friendly_tool) are stable English keys the renderers match on — never localize them. The README is English-primary (README.md) with a Chinese translation kept in sync at README.zh-CN.md; update both when either changes.
Commands
npm install # one-time; installs @tauri-apps/cli
npm run dev # tauri dev — compiles Rust and launches with hot-reloaded frontend
npm run build # tauri build — production bundles (macOS: run `npm run tauri icon app-icon.png` first for icns)
Rust-only checks (faster than a full dev launch): cargo check / cargo clippy inside src-tauri/.
There are no tests and no linter configured. The frontend is plain HTML/CSS/JS with no build step — src/ is served directly (frontendDist: "../src" in tauri.conf.json), so frontend edits need no compilation, just a reload (npm run dev hot-reloads it).
CI (.github/workflows/build.yml) builds macOS (universal) and Windows bundles via workflow_dispatch or on v* tags (tags also cut a GitHub Release). Linux isn't built in CI.
Architecture
Two halves connected by Tauri events and commands:
Rust backend (src-tauri/src/) — owns all state, runs three background threads spawned in main.rs's setup:
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.
- 5d ago First seen · 70 lines · 3,744 tokens per session scan A 4d2526ae87d3
tokibean AGENTS.md is an instructions file published in the GitHub repository ZGhey/tokibean (2 stars, last pushed 1mo ago), licensed MIT. It adds 3,744 tokens to every session, about $0.0187 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
openpets AGENTS.md
Instructions for alvinunreal/openpets, covering repository map, documentation map, tests must protect behavior, catalog direction and forward-only product direction.
clawd-on-desk CLAUDE.md
Claude Code instructions for rullerzhou-afk/clawd-on-desk, a project described as: A pixel desktop pet that watches Claude Code, Codex, Cursor & other AI coding agents — so you don't have to.
Artemis AGENTS.md
AGENTS.md instructions for momori777/Artemis, covering agents.md — tool mode, rules, capabilities (see tools.md for details), switch to roleplay and deep reasoning.
humhum CLAUDE.md
Claude Code instructions for edible999999999-jpg/humhum, covering claude.md, 项目简介, 常用命令, 架构 and 双进程模型（tauri v2）.
humhum AGENTS.md
AGENTS.md instructions for edible999999999-jpg/humhum, covering agents.md, product position, core principle, product roles and humi.
clawd-on-desk AGENTS.md
AGENTS.md instructions for rullerzhou-afk/clawd-on-desk, covering agents.md, project overview, common commands, read these docs and runtime summary.