@kairyou/agent-tools 0.3.0 → 0.4.0

package/README.md

@@ -1,28 +1,10 @@
 # Agent Tools
 
-Reusable Agent Skills, plus runtime integrations for Codex, Claude Code, and opencode.
+Reusable Agent Skills and installable integrations — statusline, provider usage, and vision — for Codex, Claude Code, and opencode.
 
-[中文](README.zh-CN.md)
-
-## Repository Structure
+Requires Node.js >= 22.
 
-```text
-agent-tools/
-├── .claude-plugin/    # Claude Code/plugin ecosystem manifest.
-├── .codex-plugin/     # Codex plugin manifest.
-├── integrations/      # Installable capabilities, one directory each; files named by agent + form.
-│   ├── statusline/    # Claude Code command-backed statusLine script.
-│   ├── usage/         # Provider usage (query core + codex hook, opencode plugins, CLI, at-usage skill template).
-│   └── vision/        # Cross-model image understanding (inspect_image MCP server + at-vision skill).
-├── skills/            # Reusable Agent Skills for CLI discovery and plugin manifests.
-│   ├── workflow/      # Workflow-oriented skills.
-│   │   ├── at-commit/   # Conventional Commit message skill.
-│   │   ├── at-review/   # Review changes for bugs and regressions.
-│   │   └── at-simplify/ # Reduce complexity and duplication in changes.
-│   └── integrations/  # Skills that integrate external systems.
-│       └── at-zentao/   # ZenTao bug/task fixing workflow.
-└── scripts/           # Install, sync, validation, and maintenance scripts.
-```
+[中文](README.zh-CN.md)
 
 ## Skills
 
@@ -91,8 +73,10 @@ Runtime capabilities, installed per agent:
 npx -y @kairyou/agent-tools@latest <capability> -a <agent...>
 ```
 
-`--dry-run` previews, `--uninstall` removes, and re-running the install command
-updates.
+`--dry-run` previews, `--uninstall` unwires the integration from the agent, and
+re-running the install command updates. The installer only touches config
+entries it wrote itself; shared files under `~/.agent-tools` are kept on
+uninstall.
 
 | Capability | Claude Code | Codex | OpenCode |
 | --- | --- | --- | --- |
@@ -122,7 +106,11 @@ it preserves top-of-file comments and existing values.
 
 ### Provider usage
 
-Shows the active API provider's balance / quota inside each agent.
+For API relay / gateway setups: shows the relay's balance / quota inside the
+agent, so when you pay per use or have plan limits you always know how much you
+have spent and how much is left — without opening the gateway console.
+Works with compatible Sub2API-like,
+NewAPI/OneAPI/OneHub/DoneHub/Veloera/AnyRouter-like, and OpenRouter gateways.
 
 ```bash
 npx -y @kairyou/agent-tools@latest usage -a claude
@@ -134,16 +122,35 @@ npx -y @kairyou/agent-tools@latest usage -a opencode
   `/at-usage` to show the current usage in the conversation.
 - **Codex** — adds a hook to `UserPromptSubmit` and `Stop` in `~/.codex/hooks.json`
   and the `at-usage` skill to `~/.agents/skills`. Run `/hooks` inside Codex once
-  to approve it. Hook output only appears in the Codex CLI; in clients that do
-  not show it (e.g. Paseo), invoke `$at-usage`.
+  to approve it. The Codex CLI displays hook output; some clients (e.g. Paseo)
+  currently do not — invoke `$at-usage` there.
 - **OpenCode** — adds server and TUI plugins: usage refreshes when the session
   goes idle and shows as a toast, and `/at-usage` shows the latest cached value.
   Restart opencode after installing or updating.
 
+The relay endpoint is auto-discovered — Codex: the active provider's `base_url`
+and key from `~/.codex/config.toml` / `auth.json`; Claude Code:
+`ANTHROPIC_BASE_URL` plus `ANTHROPIC_AUTH_TOKEN` / `ANTHROPIC_API_KEY`.
+Official (non-relay) endpoints are skipped. If it reports
+`Provider usage is unavailable.`, point it at the relay explicitly —
+`PROVIDER_USAGE_BASE_URL` / `PROVIDER_USAGE_API_KEY` env vars override the
+endpoint and key — and tune `providerUsage` in `~/.agent-tools/config.jsonc`:
+
+```jsonc
+{
+  "providerUsage": {
+    "preset": "auto", // sub2api | new-api | veloera | anyrouter | openrouter | ...
+    "userId": "",     // some NewAPI/Veloera panels require your panel user id
+    "days": 30,       // spend window for the "30d" field (max 90)
+    "debug": false    // true: log probes to ~/.agent-tools/logs/usage-debug.log
+  }
+}
+```
+
 Output examples:
 
 ```text
-# Subscription / plan quota.
+# Relay plan quota.
 API | D $0.0/$100 | W $0.0/$300 | Exp 07-08
 
 # Wallet balance.
@@ -153,11 +160,6 @@ API | balance $362 | today $61.7 | 30d $566
 Fields: `D/W/M` are daily/weekly/monthly spend against plan limits; `Exp` is
 the plan expiry; `balance` is wallet credit; `today` and `30d` are API spend.
 
-#### Supported gateways
-
-Balance, quota, and plan usage queries support compatible Sub2API-like,
-NewAPI/OneAPI/OneHub/DoneHub/Veloera/AnyRouter-like, and OpenRouter gateways.
-
 ### Vision (cross-model image understanding)
 
 Lets a main model that cannot see images ask a multimodal model specific questions about an image (local path or http(s) URL) and reason on from the answers. Typical uses: reading error screenshots, implementing UI from design mockups, locating the glitch in a bug-report screenshot. One installer capability bundling three parts: the `inspect_image` MCP stdio server, the `at-vision` policy skill, and a human diagnostic CLI.
@@ -185,8 +187,8 @@ skill into the agent's skills directory.
     "baseUrl": "https://gateway.example.com/v1",  // anthropic-compatible: gateway root, /v1/messages is appended
     "model": "internal-vlm",
     "apiKey": { "env": "OPENAI_API_KEY" }  // reuse an existing env var, or the key itself
-    // optional: "timeoutMs": 30000, "maxImageBytes": 20971520, "maxOutputTokens": 8192,
-    //           "maxConcurrentRequests": 2, "maxRequestsPerMinute": 30
+    // , "timeoutMs": 30000, "maxImageBytes": 20971520, "maxOutputTokens": 8192
+    // , "maxConcurrentRequests": 2, "maxRequestsPerMinute": 30
   }
 }
 ```
@@ -214,6 +216,26 @@ To run directly from the repository, replace the npm package name with
 npx -y github:kairyou/agent-tools usage -a codex
 ```
 
+## Repository Structure
+
+```text
+agent-tools/
+├── .claude-plugin/    # Claude Code/plugin ecosystem manifest.
+├── .codex-plugin/     # Codex plugin manifest.
+├── integrations/      # Installable capabilities, one directory each.
+│   ├── statusline/    # Agent status line: branch, model, usage.
+│   ├── usage/         # Provider balance / quota display.
+│   └── vision/        # Cross-model image understanding.
+├── skills/            # Reusable Agent Skills.
+│   ├── workflow/      # Workflow-oriented skills.
+│   │   ├── at-commit/   # Conventional Commit message skill.
+│   │   ├── at-review/   # Review changes for bugs and regressions.
+│   │   └── at-simplify/ # Reduce complexity and duplication in changes.
+│   └── integrations/  # Skills that integrate external systems.
+│       └── at-zentao/   # ZenTao bug/task fixing workflow.
+└── scripts/           # Install, sync, validation, and maintenance scripts.
+```
+
 ## FAQ
 
 ### Why does global installation fail for PromptScript?
@@ -227,8 +249,3 @@ other agents and can be ignored. See [`skills` issue #1352](https://github.com/v
 - [OpenCommit](https://github.com/di-sukharev/opencommit)
 - [GitLens](https://github.com/gitkraken/vscode-gitlens)
 - [claude-code-system-prompts](https://github.com/Piebald-AI/claude-code-system-prompts)
-
-## Notes
-
-- The installer marks and removes only the config entries it owns.
-- Run local checks with `npm test`.

package/README.zh-CN.md

@@ -1,26 +1,8 @@
 # Agent Tools
 
-可复用的 Agent Skills, 以及面向 Codex、Claude Code 与 opencode 的 runtime integrations.
+可复用的 Agent Skills, 以及面向 Codex, Claude Code 与 opencode 的可安装 integrations: statusline, provider usage 与跨模型识图.
 
-## 仓库结构
-
-```text
-agent-tools/
-├── .claude-plugin/    # Claude Code/plugin 生态的 manifest。
-├── .codex-plugin/     # Codex plugin manifest。
-├── integrations/      # 可安装的 capability, 一个一目录; 文件按 agent + 形态命名.
-│   ├── statusline/    # Claude Code command-backed statusLine 脚本.
-│   ├── usage/         # Provider usage (查询核心 + codex hook, opencode plugins, CLI, at-usage skill 模板).
-│   └── vision/        # 跨模型识图(inspect_image MCP server + 内置 at-vision skill)。
-├── skills/            # 可复用 Agent Skills，供 CLI 扫描和 plugin manifest 声明。
-│   ├── workflow/      # 工作流类 skills。
-│   │   ├── at-commit/   # 生成 Conventional Commits message.
-│   │   ├── at-review/   # 审查改动中的 bug 与回归风险.
-│   │   └── at-simplify/ # 减少改动中的冗余和复杂度.
-│   └── integrations/  # 对接外部系统的 skills。
-│       └── at-zentao/   # 禅道 bug/task 修复工作流.
-└── scripts/           # 安装、同步、校验和仓库维护脚本。
-```
+需要 Node.js >= 22.
 
 ## Skills
 
@@ -89,7 +71,8 @@ Runtime capability, 按 agent 选装:
 npx -y @kairyou/agent-tools@latest <capability> -a <agent...>
 ```
 
-`--dry-run` 预览, `--uninstall` 卸载, 重新执行安装命令即为更新.
+`--dry-run` 预览, `--uninstall` 解除对应集成, 重新执行安装命令即为更新.
+安装器只改动自己写入的配置项; 卸载会保留 `~/.agent-tools` 下的共享文件.
 
 | Capability | Claude Code | Codex | OpenCode |
 | --- | --- | --- | --- |
@@ -117,7 +100,10 @@ npx -y @kairyou/agent-tools@latest statusline -a claude
 
 ### Provider usage
 
-在各 agent 内显示当前 API provider 的余额/额度.
+面向使用 API 中转的场景: 在 agent 内直接显示中转网关的余额/额度, 按量付费或
+有套餐限额时, 随时知道花了多少, 还剩多少, 不用切出去登录网关后台.
+支持兼容 Sub2API-like, NewAPI/OneAPI/OneHub/DoneHub/Veloera/AnyRouter-like
+与 OpenRouter 的网关.
 
 ```bash
 npx -y @kairyou/agent-tools@latest usage -a claude
@@ -129,15 +115,33 @@ npx -y @kairyou/agent-tools@latest usage -a opencode
   `/at-usage` 即可在对话中显示当前用量.
 - **Codex** — 把 hook 添加到 `~/.codex/hooks.json` 的 `UserPromptSubmit` 和
   `Stop`, 并把 `at-usage` skill 装入 `~/.agents/skills`. 安装后在 Codex 里运行
-  `/hooks` 批准一次. Hook 输出只在 Codex CLI 里显示; Paseo 等不显示 hook 输出的
-  前端, 可调用 `$at-usage` 按需查询.
+  `/hooks` 批准一次. Codex CLI 会显示 hook 输出; 部分客户端(如 Paseo)目前不显示,
+  可调用 `$at-usage` 按需查询.
 - **OpenCode** — 添加 server plugin 和 TUI plugin: session 进入 idle 后自动刷新
   usage 并以 toast 显示, `/at-usage` 查看最新缓存值. 安装或更新后需要重启 opencode.
 
+中转端点自动发现 — Codex: 读取 `~/.codex/config.toml` / `auth.json` 里当前
+provider 的 `base_url` 和密钥; Claude Code: 读取 `ANTHROPIC_BASE_URL` 与
+`ANTHROPIC_AUTH_TOKEN` / `ANTHROPIC_API_KEY`. 官方(非中转)端点会被跳过.
+如果结果是 `Provider usage is unavailable.`, 可用环境变量
+`PROVIDER_USAGE_BASE_URL` / `PROVIDER_USAGE_API_KEY` 显式指定端点和密钥,
+并在 `~/.agent-tools/config.jsonc` 里调整 `providerUsage`:
+
+```jsonc
+{
+  "providerUsage": {
+    "preset": "auto", // sub2api | new-api | veloera | anyrouter | openrouter | ...
+    "userId": "",     // 部分 NewAPI/Veloera 面板需要填面板用户 id
+    "days": 30,       // "30d" 字段的统计窗口(最大 90)
+    "debug": false    // true: 探测过程写入 ~/.agent-tools/logs/usage-debug.log
+  }
+}
+```
+
 显示效果示例:
 
 ```text
-# 订阅/套餐额度.
+# 中转套餐额度.
 API | D $0.0/$100 | W $0.0/$300 | Exp 07-08
 
 # 钱包余额.
@@ -147,11 +151,6 @@ API | balance $362 | today $61.7 | 30d $566
 字段含义: `D/W/M` 是日/周/月套餐消耗与上限, `Exp` 是套餐到期日,
 `balance` 是钱包余额, `today` / `30d` 是今日与近 30 天 API 消耗.
 
-#### 支持的网关
-
-余额, 额度和套餐用量查询支持兼容 Sub2API-like,
-NewAPI/OneAPI/OneHub/DoneHub/Veloera/AnyRouter-like 与 OpenRouter 的网关.
-
 ### Vision(跨模型识图)
 
 让不支持图片的主模型借助多模态模型识图: 针对图片(本地路径或 http(s) URL)提出具体问题, 拿到答案后继续自己的推理. 常见场景: 读取报错截图, 按设计稿还原 UI, 定位测试反馈截图里的界面问题. 一个安装器 capability 包含三部分: `inspect_image` MCP stdio server, `at-vision` 策略 skill, 以及人工诊断 CLI.
@@ -178,8 +177,8 @@ OpenCode: `opencode.json`), 并把 `at-vision` skill 装入对应 agent 的 skil
     "baseUrl": "https://gateway.example.com/v1",  // anthropic-compatible 填网关根地址, 会自动拼 /v1/messages
     "model": "internal-vlm",
     "apiKey": { "env": "OPENAI_API_KEY" }  // 引用已有环境变量, 也可以直接填密钥
-    // 可选: "timeoutMs": 30000, "maxImageBytes": 20971520, "maxOutputTokens": 8192,
-    //       "maxConcurrentRequests": 2, "maxRequestsPerMinute": 30
+    // , "timeoutMs": 30000, "maxImageBytes": 20971520, "maxOutputTokens": 8192
+    // , "maxConcurrentRequests": 2, "maxRequestsPerMinute": 30
   }
 }
 ```
@@ -200,26 +199,41 @@ npx -y @kairyou/agent-tools@latest inspect-image <path|url> -q "导航栏的背
 
 ## 从 Git 运行
 
-如需直接使用仓库版本，可将 npm 包名替换为 `github:kairyou/agent-tools`（本机需已安装 Git）：
+如需直接使用仓库版本, 可将 npm 包名替换为 `github:kairyou/agent-tools` (本机需已安装 Git):
 
 ```bash
 npx -y github:kairyou/agent-tools usage -a codex
 ```
 
+## 仓库结构
+
+```text
+agent-tools/
+├── .claude-plugin/    # Claude Code/plugin 生态的 manifest.
+├── .codex-plugin/     # Codex plugin manifest.
+├── integrations/      # 可安装的 capability, 一个一目录.
+│   ├── statusline/    # Agent 状态栏: 分支, 模型, 用量.
+│   ├── usage/         # Provider 余额/额度显示.
+│   └── vision/        # 跨模型识图.
+├── skills/            # 可复用的 Agent Skills.
+│   ├── workflow/      # 工作流类 skills.
+│   │   ├── at-commit/   # 生成 Conventional Commits message.
+│   │   ├── at-review/   # 审查改动中的 bug 与回归风险.
+│   │   └── at-simplify/ # 减少改动中的冗余和复杂度.
+│   └── integrations/  # 对接外部系统的 skills.
+│       └── at-zentao/   # 禅道 bug/task 修复工作流.
+└── scripts/           # 安装, 同步, 校验和仓库维护脚本.
+```
+
 ## FAQ
 
-### 为什么全局安装提示 PromptScript 不支持？
+### 为什么全局安装提示 PromptScript 不支持?
 
 `PromptScript does not support global skill installation` 表示 PromptScript Agent 不支持全局
-安装，不影响其他 Agent，可忽略。参见 [`skills` issue #1352](https://github.com/vercel-labs/skills/issues/1352)。
+安装, 不影响其他 Agent, 可忽略. 参见 [`skills` issue #1352](https://github.com/vercel-labs/skills/issues/1352).
 
 ## References
 
 - [OpenCommit](https://github.com/di-sukharev/opencommit)
 - [GitLens](https://github.com/gitkraken/vscode-gitlens)
 - [claude-code-system-prompts](https://github.com/Piebald-AI/claude-code-system-prompts)
-
-## 说明
-
-- 安装器只标记并移除自己写入的配置项。
-- 本地检查运行 `npm test`。

package/dist/statusline/claude-statusline.mjs

@@ -0,0 +1,355 @@
+#!/usr/bin/env node
+
+// integrations/statusline/claude-statusline.mjs
+import { execFileSync, spawn } from "node:child_process";
+import fs from "node:fs";
+import { basename, dirname, join } from "node:path";
+import { fileURLToPath } from "node:url";
+var SCRIPT_DIR = dirname(fileURLToPath(import.meta.url));
+var AGENT_TOOLS_HOME = process.env.AGENT_TOOLS_HOME || join(SCRIPT_DIR, "..", "..");
+var DEFAULT_CONFIG_FILE = join(AGENT_TOOLS_HOME, "config.jsonc");
+var SNAPSHOT_FILE = join(AGENT_TOOLS_HOME, "cache", "usage-snapshot.json");
+var REFRESH_STATE_FILE = join(AGENT_TOOLS_HOME, "cache", "usage-refresh-state.json");
+var USAGE_RUNTIME = join(AGENT_TOOLS_HOME, "dist", "usage", "core.mjs");
+var DEFAULT_SNAPSHOT_TTL_MS = 6e4;
+var DEFAULT_REFRESH_COOLDOWN_MS = 3e4;
+var DEFAULT_FAILURE_BACKOFF_MS = 12e4;
+var DEFAULT_CONFIG = {
+  fields: ["branch", "model", "fiveHour", "week"],
+  separator: " | ",
+  symbols: {
+    branch: "\u2387",
+    reset: "\u27F3",
+    empty: "\u2013",
+    fiveHour: "5h",
+    week: "w",
+    context: "ctx"
+  }
+};
+var FIELD_ALIASES = {
+  cwd: "directory",
+  dir: "directory",
+  five: "fiveHour",
+  five_hour: "fiveHour",
+  "5h": "fiveHour",
+  sevenDay: "week",
+  seven_day: "week",
+  "7d": "week",
+  weekly: "week",
+  ctx: "context"
+};
+async function readStdin() {
+  const chunks = [];
+  for await (const chunk of process.stdin) chunks.push(chunk);
+  return Buffer.concat(chunks).toString("utf8");
+}
+function readJsonFile(file) {
+  try {
+    if (!fs.existsSync(file)) return {};
+    const raw = stripTrailingCommas(stripJsonComments(fs.readFileSync(file, "utf8").replace(/^\uFEFF/, "")));
+    return raw.trim() ? JSON.parse(raw) : {};
+  } catch {
+    return {};
+  }
+}
+function stripJsonComments(input) {
+  let out = "";
+  let inString = false;
+  let escaped = false;
+  for (let i = 0; i < input.length; i++) {
+    const ch = input[i];
+    const next = input[i + 1];
+    if (inString) {
+      out += ch;
+      escaped = ch === "\\" ? !escaped : false;
+      if (ch === '"' && !escaped) inString = false;
+      continue;
+    }
+    if (ch === '"') {
+      inString = true;
+      out += ch;
+      continue;
+    }
+    if (ch === "/" && next === "/") {
+      while (i < input.length && input[i] !== "\n") i++;
+      out += "\n";
+      continue;
+    }
+    if (ch === "/" && next === "*") {
+      i += 2;
+      while (i < input.length && !(input[i] === "*" && input[i + 1] === "/")) i++;
+      i++;
+      continue;
+    }
+    out += ch;
+  }
+  return out;
+}
+function stripTrailingCommas(input) {
+  let out = "";
+  let inString = false;
+  let escaped = false;
+  for (let i = 0; i < input.length; i++) {
+    const ch = input[i];
+    if (inString) {
+      out += ch;
+      escaped = ch === "\\" ? !escaped : false;
+      if (ch === '"' && !escaped) inString = false;
+      continue;
+    }
+    if (ch === '"') {
+      inString = true;
+      out += ch;
+      continue;
+    }
+    if (ch === ",") {
+      let j = i + 1;
+      while (j < input.length && /\s/.test(input[j])) j++;
+      if (input[j] === "}" || input[j] === "]") continue;
+    }
+    out += ch;
+  }
+  return out;
+}
+function parseArgs(argv) {
+  const opts = {};
+  for (let i = 0; i < argv.length; i++) {
+    const arg = argv[i];
+    if (arg === "--fields" && argv[i + 1]) {
+      opts.fields = argv[++i];
+    } else if (arg.startsWith("--fields=")) {
+      opts.fields = arg.slice("--fields=".length);
+    } else if (arg === "--separator" && argv[i + 1]) {
+      opts.separator = argv[++i];
+    } else if (arg.startsWith("--separator=")) {
+      opts.separator = arg.slice("--separator=".length);
+    }
+  }
+  return opts;
+}
+function splitFields(value) {
+  if (Array.isArray(value)) return value;
+  if (typeof value !== "string") return null;
+  return value.split(/[,\s]+/).map((part) => part.trim()).filter(Boolean);
+}
+function normalizeField(field) {
+  return FIELD_ALIASES[field] || field;
+}
+function mergeConfig(cli) {
+  const rootConfig = readJsonFile(DEFAULT_CONFIG_FILE);
+  const fileConfig = rootConfig.statusline || {};
+  const envFields = process.env.AGENT_TOOLS_STATUSLINE_FIELDS;
+  const envSeparator = process.env.AGENT_TOOLS_STATUSLINE_SEPARATOR;
+  const config = {
+    ...DEFAULT_CONFIG,
+    ...fileConfig,
+    symbols: { ...DEFAULT_CONFIG.symbols, ...fileConfig.symbols || {} }
+  };
+  const fields = splitFields(cli.fields) || splitFields(envFields) || splitFields(fileConfig.fields) || DEFAULT_CONFIG.fields;
+  config.fields = fields.map(normalizeField);
+  if (typeof envSeparator === "string") config.separator = envSeparator;
+  if (typeof cli.separator === "string") config.separator = cli.separator;
+  return config;
+}
+function gitBranch(cwd) {
+  try {
+    const out = execFileSync("git", ["rev-parse", "--abbrev-ref", "HEAD"], {
+      cwd,
+      stdio: ["ignore", "pipe", "ignore"],
+      encoding: "utf8"
+    }).trim();
+    return out && out !== "HEAD" ? out : "";
+  } catch {
+    return "";
+  }
+}
+function secondsUntil(unixSeconds) {
+  if (!unixSeconds) return null;
+  const seconds = Number(unixSeconds) - Math.floor(Date.now() / 1e3);
+  return Number.isFinite(seconds) ? Math.max(0, seconds) : null;
+}
+function compactDuration(totalSeconds) {
+  if (totalSeconds == null) return "";
+  if (totalSeconds <= 0) return "0m";
+  let seconds = totalSeconds;
+  const days = Math.floor(seconds / 86400);
+  seconds %= 86400;
+  const hours = Math.floor(seconds / 3600);
+  seconds %= 3600;
+  const minutes = Math.floor(seconds / 60);
+  if (days) return `${days}d${hours}h`;
+  if (hours) return `${hours}h${minutes}m`;
+  return `${minutes}m`;
+}
+function usageWindow(window, config) {
+  if (!window || typeof window.used_percentage !== "number") {
+    return "";
+  }
+  const pct = `${Math.round(window.used_percentage)}%`;
+  const left = compactDuration(secondsUntil(window.resets_at));
+  return left ? `${pct} ${config.symbols.reset}${left}` : pct;
+}
+function showMissingUsageWindow() {
+  const baseUrl = activeRelayBaseUrl();
+  return !baseUrl || isOfficialBaseUrl(baseUrl);
+}
+function shortModelName(name) {
+  if (!name) return "";
+  return String(name).replace(/^Claude\s+/i, "").replace(/\s*\[1m\]\s*$/i, "").trim();
+}
+function renderField(field, data, config) {
+  const dir = data?.workspace?.current_dir || data?.cwd || process.cwd() || "";
+  const projectDir = data?.workspace?.project_dir || dir;
+  switch (field) {
+    case "branch": {
+      const branch = gitBranch(projectDir);
+      return branch ? `${config.symbols.branch} ${branch}` : "";
+    }
+    case "model":
+      return shortModelName(data?.model?.display_name || data?.model?.id || "");
+    case "fiveHour": {
+      const value = usageWindow(data?.rate_limits?.five_hour, config);
+      return value || showMissingUsageWindow() ? `${config.symbols.fiveHour} ${value || config.symbols.empty}` : "";
+    }
+    case "week": {
+      const value = usageWindow(data?.rate_limits?.seven_day, config);
+      return value || showMissingUsageWindow() ? `${config.symbols.week} ${value || config.symbols.empty}` : "";
+    }
+    case "context": {
+      const pct = data?.context_window?.used_percentage;
+      return typeof pct === "number" ? `${config.symbols.context} ${Math.round(pct)}%` : "";
+    }
+    case "directory":
+      return dir ? basename(dir) : "";
+    default:
+      return "";
+  }
+}
+function numberFromEnv(name, fallback) {
+  const value = Number(process.env[name]);
+  return Number.isFinite(value) && value >= 0 ? value : fallback;
+}
+function cleanBaseUrl(baseUrl) {
+  return String(baseUrl || "").replace(/\/+$/, "");
+}
+function isOfficialBaseUrl(baseUrl) {
+  if (!baseUrl) return true;
+  const clean = cleanBaseUrl(baseUrl);
+  return [
+    "https://api.anthropic.com",
+    "https://api.anthropic.com/v1",
+    "https://api.openai.com",
+    "https://api.openai.com/v1"
+  ].includes(clean);
+}
+function usageRouteCacheKey(baseUrl) {
+  try {
+    const url = new URL(cleanBaseUrl(baseUrl));
+    url.hash = "";
+    url.search = "";
+    url.pathname = url.pathname.replace(/\/+$/, "").replace(/\/api\/v1$/i, "").replace(/\/v1$/i, "");
+    return url.toString().replace(/\/$/, "");
+  } catch {
+    return cleanBaseUrl(baseUrl).endsWith("/v1") ? cleanBaseUrl(baseUrl).slice(0, -3) : cleanBaseUrl(baseUrl);
+  }
+}
+function readJsonFileRaw(file) {
+  try {
+    if (!fs.existsSync(file)) return {};
+    const raw = fs.readFileSync(file, "utf8").replace(/^\uFEFF/, "");
+    return raw.trim() ? JSON.parse(raw) : {};
+  } catch {
+    return {};
+  }
+}
+function activeRelayBaseUrl() {
+  return process.env.PROVIDER_USAGE_BASE_URL || process.env.ANTHROPIC_BASE_URL || "";
+}
+function hasClaudeUsageToken() {
+  return Boolean(
+    process.env.PROVIDER_USAGE_API_KEY || process.env.ANTHROPIC_AUTH_TOKEN || process.env.ANTHROPIC_API_KEY
+  );
+}
+function snapshotForBaseUrl(baseUrl) {
+  const snapshot = readJsonFileRaw(SNAPSHOT_FILE);
+  const key = usageRouteCacheKey(baseUrl);
+  const item = snapshot?.items?.[key];
+  return item?.text ? item : null;
+}
+function refreshStateForBaseUrl(baseUrl) {
+  const state = readJsonFileRaw(REFRESH_STATE_FILE);
+  return state?.items?.[usageRouteCacheKey(baseUrl)] || {};
+}
+function ageMs(isoDate) {
+  const time = Date.parse(isoDate || "");
+  return Number.isFinite(time) ? Date.now() - time : Number.POSITIVE_INFINITY;
+}
+function writeRefreshAttempt(baseUrl) {
+  try {
+    const state = readJsonFileRaw(REFRESH_STATE_FILE);
+    const key = usageRouteCacheKey(baseUrl);
+    state.version = 1;
+    state.items = state.items && typeof state.items === "object" ? state.items : {};
+    state.items[key] = {
+      ...state.items[key] || {},
+      baseUrl,
+      lastAttemptAt: (/* @__PURE__ */ new Date()).toISOString()
+    };
+    fs.mkdirSync(dirname(REFRESH_STATE_FILE), { recursive: true });
+    fs.writeFileSync(REFRESH_STATE_FILE, `${JSON.stringify(state, null, 2)}
+`);
+  } catch {
+  }
+}
+function shouldRefreshUsage(baseUrl, snapshot) {
+  if (process.env.AGENT_TOOLS_USAGE_REFRESH === "0") return false;
+  if (!baseUrl || isOfficialBaseUrl(baseUrl)) return false;
+  if (!hasClaudeUsageToken()) return false;
+  if (!fs.existsSync(USAGE_RUNTIME)) return false;
+  const ttlMs = numberFromEnv("AGENT_TOOLS_USAGE_SNAPSHOT_TTL_MS", DEFAULT_SNAPSHOT_TTL_MS);
+  const cooldownMs = numberFromEnv("AGENT_TOOLS_USAGE_REFRESH_COOLDOWN_MS", DEFAULT_REFRESH_COOLDOWN_MS);
+  const failureBackoffMs = numberFromEnv("AGENT_TOOLS_USAGE_FAILURE_BACKOFF_MS", DEFAULT_FAILURE_BACKOFF_MS);
+  const state = refreshStateForBaseUrl(baseUrl);
+  if (ageMs(state.lastAttemptAt) < cooldownMs) return false;
+  if (state.lastError && ageMs(state.lastFailureAt) < failureBackoffMs) return false;
+  return !snapshot || ageMs(snapshot.updatedAt) >= ttlMs;
+}
+function refreshUsageInBackground(baseUrl) {
+  if (!shouldRefreshUsage(baseUrl, snapshotForBaseUrl(baseUrl))) return;
+  writeRefreshAttempt(baseUrl);
+  try {
+    const child = spawn(process.execPath, [USAGE_RUNTIME, "refresh", "--agent", "claude"], {
+      detached: true,
+      stdio: "ignore",
+      env: process.env,
+      windowsHide: true
+    });
+    child.unref();
+  } catch {
+  }
+}
+function providerUsageStatus() {
+  const baseUrl = activeRelayBaseUrl();
+  if (!baseUrl || isOfficialBaseUrl(baseUrl)) return "";
+  const snapshot = snapshotForBaseUrl(baseUrl);
+  if (shouldRefreshUsage(baseUrl, snapshot)) refreshUsageInBackground(baseUrl);
+  return snapshot?.text || "";
+}
+function render(data, config) {
+  const fields = config.fields.map((field) => renderField(field, data, config)).filter(Boolean);
+  const providerUsage = providerUsageStatus();
+  if (providerUsage) fields.push(providerUsage);
+  return fields.join(config.separator);
+}
+async function main() {
+  const config = mergeConfig(parseArgs(process.argv.slice(2)));
+  let data = {};
+  try {
+    const raw = await readStdin();
+    data = raw.trim() ? JSON.parse(raw) : {};
+  } catch {
+    data = {};
+  }
+  process.stdout.write(render(data, config));
+}
+main();