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/viganogabriele/agent-usage-plus/agents-mdgit clone --depth 1 https://github.com/viganogabriele/agent-usage-plusWrote 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/viganogabriele/agent-usage-plus/agents-md)<a href="https://agentmods.dev/instructions/viganogabriele/agent-usage-plus/agents-md"><img src="https://agentmods.dev/badge/instructions/viganogabriele/agent-usage-plus/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 | $0.01622 | $0.01622 |
| Opus 5 | $0.00811 | $0.00811 |
| Sonnet 5 | $0.00324 | $0.00324 |
| Haiku 4.5 | $0.00162 | $0.00162 |
Grade A, and why
agent-usage-plus AGENTS.md scanned grade A with 0 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 3d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 125 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Usage Plus
Omarchy/Quickshell bar widget for local AI-subscription usage. Keep changes small, data-safe, and visually verified; a plausible QML diff is not enough.
Map
Agent.qml: widget entry point and plugin metadata bridge.Main.qml: settings, records, refresh and bar-layout state.Panel.qml: bar and panel UI.logic/: pure JS, covered by Node tests intest/.collectors/: bundled collectors, with their own Python tests incollectors/tests/. Readcollectors/README.mdanddocs/collector-contract.mdbefore changing their output.
Working rules
- Never read or write API keys, OAuth tokens, transcripts, or shell settings
directly. Collectors emit records; settings changes go through the existing
spawned
omarchy bar setpath. - Preserve user worktree changes. Do not push or open PRs unless explicitly requested.
- Add or adjust a test in
logic/for logic changes. Keep collector output compliant withdocs/collector-contract.md, including useful auth-missing and endpoint-down states. - Use bundled SVGs for provider marks. Set
sourceSizeon every markImage(width * Screen.devicePixelRatio, matching Tray.qml/Menu.qml elsewhere in the shell) — without it marks read as soft/blurry, worse the smaller the box or the higher the output's pixel density. Do not replace SVGs with scaled raster assets; check bar-scale and panel-scale rendering separately. - Follow the theme contract: surfaces, foregrounds, tracks, fonts, and the
critical state come from Omarchy's live colors; Warn is intentionally fixed
amber (
#F2B705) for a stable, distinct warning state. Bar marks must choose their default/light asset from the bar's live foreground, including hover or transparent-bar states, so the bar and panel stay visually consistent. - Treat provider responses as untrusted too: keep collector JSON reads bounded before parsing, preserve explicit auth/endpoint error states, and never put an upstream response body or credential into a record or log.
omarchy bar set <id> <key> <value> --jsoncannot take a top-level JSON array value with more than one element — it miscounts its own arguments and fails every time ("Too many arguments provided"), silently, with no QML-side error beyond aconsole.warnline in the shell's own log (nothing surfaces in the panel). A single-element array, and any JSON object no matter how many keys, both write fine — only a bare multi-element array is affected. Reproduce/verify directly:omarchy bar set <id> <key> '["a","b"]' --jsonfails,omarchy bar set <id> <key> '"[\"a\",\"b\"]"' --json(the array double-JSON-encoded into a string) succeeds. Any new array-valued setting must be written double-encoded (JSON.stringify(JSON.stringify(arr))) and unwrapped with a guardedJSON.parseon read — seeproviderOrderin Main.qml for the pattern.- Do not add a hardcoded cap on how many providers can show in the bar (no
"3 fixed + cycle slots" style budget). The only real ceiling is however
many providers the person has actually configured (Fixed count + Cycle
slots), which is itself already bounded by how many providers exist to
configure.
selectBarLayout's ownMath.min(10, ...)— the number of bundled collectors — is the one non-arbitrary safety clamp;barSlotLimit(Main.qml) andmaxBarProviderSlots(Panel.qml) exist only soselectBarLayout/the "+N" affordance have a number to work with and should stay far above anything a real configuration would hit. - A notification not appearing is not necessarily a code bug: Omarchy's own
Do Not Disturb state lives in
~/.local/state/omarchy/notifications.json("dnd": truesuppresses the popup even thoughnotify-sendstill succeeds), and every notification actually delivered is logged under~/.local/state/omarchy/notifications/history/*.jsonregardless of DND — check that history before assuming the panel's own notify-send dispatch is broken. - Reordering draggable items (the panel's provider switcher) does not move a
real
Grid/Row/Columnchild and rely on the positioner "leaving a Drag.active item alone" to snap it back afterward — that did not reliably reposition the item once dropped, leaving marks stranded wherever they were released. Use a separate, non-positioned ghost item (parented at the panel's top level, positioned viamapToItem) that follows the pointer; the real grid children never move and the model is only reordered once, on drop — seeproviderSwitch/dragGhostin Panel.qml.
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.
- 3d ago First seen · 125 lines · 1,622 tokens per session scan A 1cfab5bdf710
agent-usage-plus AGENTS.md is an instructions file published in the GitHub repository viganogabriele/agent-usage-plus (2 stars, last pushed 4d ago), licensed MIT. It adds 1,622 tokens to every session, about $0.0081 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.