繁體中文 | English
A multilingual, cross-platform skill that customizes the Antigravity CLI statusline (footer) — pick exactly which indicators to show, in any order you like, with smart line wrapping out of the box.
| English (us) | Traditional Chinese (zh-tw) | Japanese (jp) |
|---|---|---|
![]() |
![]() |
![]() |
| English (us) | Traditional Chinese (zh-tw) | Japanese (jp) |
|---|---|---|
![]() |
![]() |
![]() |
- Node.js (required) — the renderer scripts are pure
.mjs. Without Node.js the statusline stays blank andagywill auto-disable it after repeated failures. The skill pre-checks this for you. - Git (optional) — needed for
git-branch,vcs-dirty, andvcs-typeindicators.
agy plugin install https://github.com/andyawd/antigravity-cli-statuslineThe CLI stages the bundle at
~/.gemini/antigravity-cli/plugins/antigravity-cli-statusline/.
In the Antigravity CLI prompt, type:
/antigravity-cli-statusline
The skill walks you through language selection, indicator picking, and sorting, then deploys the renderer scripts and writes the three-layer settings.json. The statusline updates live without a CLI restart.
AI Model & Agent
- Current AI model name (
model-name) — the model serving the current conversation - Active agent profile (
agent-profile) — the loaded Agent Profile name - Agent state (
agent-state) —idle / thinking / working / tool_use / initializing - AI credits (
ai-credits) — remaining account AI credits
Quota & Tokens
- Account API available quota (
quota) — percentage, color-coded across four tiers - API reset countdown (
quota-reset-countdown) — time remaining until quota refreshes - Context window usage (
context-used) — percentage of context consumed - Session token count (
token-count) — precise token usage this session - Cumulative AI artifacts (
artifacts) — files/outputs produced this session - Account plan tier (
plan-tier) — current subscription level
Interactive State
- Pending tool confirmation (
tool-confirmation) — a tool dialog is waiting on you - Pending user input queue (
pending-input) — queued inputs waiting to run - Running background tasks (
background-tasks) — active background task count - Active subagents (
subagents) — running subagent count
Project & VCS
- Project short path (
project-path) — current workspace basename - Project full path (
project-full-path) — absolute workspace path - VCS type (
vcs-type) —git / jj / fig - Current Git branch (
git-branch) - Working tree status (
vcs-dirty) —dirty / clean
System & Account
- CLI RAM usage (
memory-usage) — RSS memory consumed by theagyprocess - Antigravity CLI version (
cli-version) - Conversation ID (
conversation-id) — first 8 chars, for debugging - Sandbox mode (
sandbox-status) —off / on (net) / on (no-net) - Account email (
account-email)
In the skill's third stage (Step 4), enter a comma-separated list of numbers in the Write-in box to set the order:
2,5,1
- Numbers refer to the position in your Step 3 selection (1-based)
- Identifiers like
quotaormodel-namealso work and can be mixed in, but numbers are the most direct - Indicators you don't list are dropped — the result is your final display set
- Leave it blank or pick
(Recommended) Skipto keep the original selection order with all indicators shown
Two mechanisms work together:
- Smart automatic wrapping — the renderer reads the terminal width and wraps the next indicator to a new line when it would overflow. No setup required.
- Manual newline — insert the
ntoken in your sort string to force a line break at that position. Reusable:
1,2,n,3,4
(If you picked 4 indicators in Step 3, this puts items 1–2 on line one and forces 3–4 onto line two.)
A 24-bit truecolor four-tier palette (Blue → Green → Yellow → Pink) shifts with API quota and context usage; each AI model family also gets its own brand accent color.
Contributions are welcome — including new indicators, cross-platform fixes, and AI-assisted language translations. See CONTRIBUTING.md for the workflow.
Special thanks to 60ke/antigravity-statusline for the inspiration behind the quota monitoring feature. Their original project paved the way for statusline integrations. Since their implementation was written in Python, which could be challenging to execute consistently across Windows and macOS, this project was rewritten in JavaScript (Node.js) to ensure seamless cross-platform compatibility without relying on external Python dependencies.
This project is licensed under the MIT License.





