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/fschmutz/claude-usage-panel/claude-mdgit clone --depth 1 https://github.com/fschmutz/claude-usage-panelWrote 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/fschmutz/claude-usage-panel/claude-md)<a href="https://agentmods.dev/instructions/fschmutz/claude-usage-panel/claude-md"><img src="https://agentmods.dev/badge/instructions/fschmutz/claude-usage-panel/claude-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 | $0.02932 | $0.02932 |
| Opus 5 | $0.01466 | $0.01466 |
| Sonnet 5 | $0.00586 | $0.00586 |
| Haiku 4.5 | $0.00293 | $0.00293 |
Grade D, and why
claude-usage-panel CLAUDE.md scanned grade D with 3 findings 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.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
# curl -fsSL https://fschmutz.github.io/claude-usage-panel/install | bash [-s -- target…]) Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
./install.sh statusline # status line → merges into ~/.claude/settings.json Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
# curl -fsSL https://fschmutz.github.io/claude-usage-panel/install | bash [-s -- target…]) How it starts
The opening of the file, as written. The whole thing — 201 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What this is
Four native clients that surface Claude Code plan-usage limits from the official
Anthropic usage endpoint: a GNOME Shell extension, a macOS SwiftUI menu-bar
app, a Node status line for under the Claude Code prompt, and an MCP
server (mcp/server.js) exposing a get_usage tool to Claude Code / Cursor
(installable as a Claude Code plugin from .claude-plugin/marketplace.json +
plugin/, or via npx -y github:fschmutz/claude-usage-panel - package.json has
a bin entry). All read a locally-stored OAuth token read-only and render the
same limits[] data. An optional Cursor team-spend section is available in the
two desktop clients.
Commands
# JS unit tests (GNOME lib + status line) - the primary test gate
npm test # node --test tests/*.test.js
node --test tests/pure.test.js # single file
node --test --test-name-pattern="sparkline" tests/pure.test.js # single test
# Swift core unit tests (runs on Linux CI too - no macOS needed)
cd macos && swift test
# Lint / format everything (same set runs in CI on every push, as the `lint` job)
pre-commit run --all-files
pre-commit run eslint --all-files # single hook
pre-commit run zizmor --all-files # workflow security audit (actionlint = validity)
# Install - one unified entrypoint for all clients (also reachable with
# curl -fsSL https://fschmutz.github.io/claude-usage-panel/install | bash [-s -- target…])
./install.sh # auto-detect OS → install the sensible set
./install.sh gnome # GNOME extension only → then log out/in (Wayland)
./install.sh statusline # status line → merges into ~/.claude/settings.json
./install.sh mcp # MCP server → claude mcp add + ~/.cursor/mcp.json
./install.sh macos # build macos/ClaudeUsagePanel.app
./install.sh autoupdate # schedule the daily update check (systemd timer / launchd / cron)
./install.sh sessionping 05:30 10:35 --days=mon-fri # scheduled claude pings that open the 5h session window (opt-in)
./install.sh update [target...] # reinstall installed targets (upgrade); --pull to git pull first
./install.sh --uninstall [target...] # reverse it | --list (detected + installed) | -h
./install.sh --dry-run [target...] # print actions without touching anything
# Screenshots are GENERATED - after any UI-visible contract change:
node scripts/screenshots/render.mjs # rewrites docs/screenshot.svg + og.svg
node scripts/screenshots/render.mjs --check # what CI runs; exits 1 on drift
# Release: bump the version everywhere from one source of truth
./scripts/bump-version.sh 1.4.0
# Daily auto-update worker (what the timer runs) - safe to run by hand
./scripts/auto-update.sh --status # installed vs newest released tag, last check
./scripts/auto-update.sh --check # check only; exit 10 = update available
# Session-ping worker (what the sessionping schedule runs) - safe to run by hand
./scripts/session-ping.sh --status # configured times/days, last ping, log path
./scripts/session-ping.sh --force # ping now, whatever the day is
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 · 201 lines · 2,932 tokens per session scan D 571455ba117e
claude-usage-panel CLAUDE.md is an instructions file published in the GitHub repository fschmutz/claude-usage-panel (4 stars, last pushed today), licensed MIT. It adds 2,932 tokens to every session, about $0.0147 per session on Opus 5. A static security scan graded it D with 3 findings (downloads and executes remote code, reads agent configuration directories, 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
ccstory AGENTS.md
Instructions for atomchung/ccstory, covering ccstory agent guide, instruction layers, repository contract, concurrent-agent safety and development and verification.
gjsify AGENTS.md
AGENTS.md instructions for gjsify/gjsify, covering agents.md — gjsify, where the rules live — nearest agents.md wins, governance — non-negotiable, structure and runtime & platform model.
ccstory CLAUDE.md
Instructions for atomchung/ccstory, a project described as: Claude Code usage recap with narrative. ccusage tells you the bill, ccstory tells the story.
tokenmaxxing AGENTS.md
Instructions for 851-labs/tokenmaxxing, covering repository instructions, effect error style, export style, conventions and cli output style.
FlyCrys AGENTS.md
AGENTS.md instructions for SergKam/FlyCrys, covering agents.md - flycrys, rules, tech stack, architecture and layered architecture (must follow).
sessions-chronicle AGENTS.md
AGENTS.md instructions for supermaciz/sessions-chronicle, covering repository guidelines, where to look first, terminology, project structure & module organization and fast dev loop.