Feat: Add Antigravity agent adapter

README.md

@@ -4,15 +4,15 @@
 
 <p align="center"><b>Live page:</b> <a href="https://open-design.ai/html-anything/"><b>open-design.ai/html-anything/</b></a> — overview, surface modes, and showcase before you clone.</p>
 
-> **Markdown is the draft. HTML is what humans read. Your local agent writes it.** The agentic HTML editor — in the agentic era, you don't hand-edit docs anymore, so the output format should be what the reader actually wants: HTML. Local-first, zero API key, reuses the CLI session you already have logged in — **9 coding-agent CLIs** auto-detected on your `PATH` (Claude Code · Cursor Agent · Codex · Gemini CLI · GitHub Copilot CLI · OpenCode · Qwen Coder · Aider · IBM Bob), driven by **75 composable skill templates** across **9 deliverable surfaces** (magazine articles · keynote decks · résumés · posters · Xiaohongshu cards · tweet cards · web prototypes · data reports · Hyperframes videos). One-click export to WeChat / X / Zhihu, or download `.html` / `.png`.
+> **Markdown is the draft. HTML is what humans read. Your local agent writes it.** The agentic HTML editor — in the agentic era, you don't hand-edit docs anymore, so the output format should be what the reader actually wants: HTML. Local-first, zero API key, reuses the CLI session you already have logged in — **10 coding-agent CLIs** auto-detected on your `PATH` (Claude Code · Cursor Agent · Codex · Gemini CLI · GitHub Copilot CLI · OpenCode · Qwen Coder · Aider · IBM Bob · Antigravity), driven by **75 composable skill templates** across **9 deliverable surfaces** (magazine articles · keynote decks · résumés · posters · Xiaohongshu cards · tweet cards · web prototypes · data reports · Hyperframes videos). One-click export to WeChat / X / Zhihu, or download `.html` / `.png`.
 
 <p align="center">
   <img src="docs/assets/banner.png" alt="HTML Anything — the agentic HTML editor, on your laptop" width="100%" />
 </p>
 
 <p align="center">
   <a href="LICENSE"><img alt="License" src="https://img.shields.io/badge/license-Apache%202.0-blue.svg?style=flat-square" /></a>
-  <a href="#supported-coding-agents"><img alt="Agents" src="https://img.shields.io/badge/agents-9%20CLIs-black?style=flat-square" /></a>
+  <a href="#supported-coding-agents"><img alt="Agents" src="https://img.shields.io/badge/agents-10%20CLIs-black?style=flat-square" /></a>
   <a href="#skills"><img alt="Skills" src="https://img.shields.io/badge/skills-75-orange?style=flat-square" /></a>
   <a href="#export-targets"><img alt="Export" src="https://img.shields.io/badge/export-WeChat%20%C2%B7%20X%20%C2%B7%20Zhihu%20%C2%B7%20PNG-9b59b6?style=flat-square" /></a>
   <a href="#quickstart"><img alt="Quickstart" src="https://img.shields.io/badge/quickstart-30%20seconds-green?style=flat-square" /></a>
@@ -203,6 +203,7 @@ On startup we scan `PATH` (including `~/.local/bin`, `~/.bun/bin`, `/opt/homebre
 | **Qwen Coder** | `qwen` | `qwen --yolo -` |
 | **Aider** | `aider` | `aider --no-pretty --no-stream --yes-always --message-file -` |
 | **IBM Bob** | `bob` | `bob --output-format stream-json --hide-intermediary-output` |
+| **Antigravity** | `agy` | `agy --output-format stream-json --dangerously-skip-permissions --disable-slash-commands --effort medium -p <prompt>` |
 
 > The detection strategy and per-CLI adapter shape are borrowed directly from [`nexu-io/open-design`](https://github.com/nexu-io/open-design) and [`multica-ai/multica`](https://github.com/multica-ai/multica): one privileged process spawns CLIs, JSON-line is the wire protocol, every CLI gets a thin adapter in [`next/src/lib/agents/argv.ts`](next/src/lib/agents/argv.ts).
 

README.zh-CN.md

@@ -4,15 +4,15 @@
 
 <p align="center"><b>项目主页:</b> <a href="https://open-design.ai/html-anything/"><b>open-design.ai/html-anything/</b></a> —— 不用 clone 也能先看看 HTML Anything 长什么样、能干啥。</p>
 
-> **Markdown 是草稿, HTML 才是给人读的成品 —— 让本地 agent 直接写 HTML。** Agent 时代的 HTML 编辑器 —— 既然你已经不亲手改文档、全都让 Claude 改了, 那 agent 的输出就该是读者真正想看的 HTML, 而不是中间态的 markdown。本地优先、零 API Key、复用你已经登录好的 CLI session —— **9 个 coding-agent CLI** 在 `PATH` 上自动识别（Claude Code · Cursor Agent · Codex · Gemini CLI · GitHub Copilot CLI · OpenCode · Qwen Coder · Aider · IBM Bob），驱动 **75 套 skill 模板** 和 **9 类可交付场景**（杂志文章 · Keynote PPT · 简历 · 海报 · 小红书 · 推特卡 · Web 原型 · 数据报告 · Hyperframes 视频）。一键复制到公众号 / 推特 / 知乎，或者下载 `.html` / `.png`。
+> **Markdown 是草稿, HTML 才是给人读的成品 —— 让本地 agent 直接写 HTML。** Agent 时代的 HTML 编辑器 —— 既然你已经不亲手改文档、全都让 Claude 改了, 那 agent 的输出就该是读者真正想看的 HTML, 而不是中间态的 markdown。本地优先、零 API Key、复用你已经登录好的 CLI session —— **10 个 coding-agent CLI** 在 `PATH` 上自动识别（Claude Code · Cursor Agent · Codex · Gemini CLI · GitHub Copilot CLI · OpenCode · Qwen Coder · Aider · IBM Bob · Antigravity），驱动 **75 套 skill 模板** 和 **9 类可交付场景**（杂志文章 · Keynote PPT · 简历 · 海报 · 小红书 · 推特卡 · Web 原型 · 数据报告 · Hyperframes 视频）。一键复制到公众号 / 推特 / 知乎，或者下载 `.html` / `.png`。
 
 <p align="center">
   <img src="docs/assets/banner.png" alt="HTML Anything — agent 时代的 HTML 编辑器，在你的笔记本上" width="100%" />
 </p>
 
 <p align="center">
   <a href="LICENSE"><img alt="License" src="https://img.shields.io/badge/license-Apache%202.0-blue.svg?style=flat-square" /></a>
-  <a href="#-自动识别本地-agent"><img alt="Agents" src="https://img.shields.io/badge/agents-9%20CLIs-black?style=flat-square" /></a>
+  <a href="#-自动识别本地-agent"><img alt="Agents" src="https://img.shields.io/badge/agents-10%20CLIs-black?style=flat-square" /></a>
   <a href="#-skills"><img alt="Skills" src="https://img.shields.io/badge/skills-75-orange?style=flat-square" /></a>
   <a href="#一键发布到平台"><img alt="Export" src="https://img.shields.io/badge/export-WeChat%20%C2%B7%20X%20%C2%B7%20Zhihu%20%C2%B7%20PNG-9b59b6?style=flat-square" /></a>
   <a href="#-30-秒上手"><img alt="Quickstart" src="https://img.shields.io/badge/quickstart-30%20seconds-green?style=flat-square" /></a>
@@ -203,6 +203,7 @@ pnpm -F @html-anything/e2e test
 | **Qwen Coder** | `qwen` | `qwen --yolo -` |
 | **Aider** | `aider` | `aider --no-pretty --no-stream --yes-always --message-file -` |
 | **IBM Bob** | `bob` | `bob --output-format stream-json --hide-intermediary-output` |
+| **Antigravity** | `agy` | `agy --output-format stream-json --dangerously-skip-permissions --disable-slash-commands --effort medium -p <prompt>` |
 
 > 这一层的设计直接借鉴了 [`nexu-io/open-design`](https://github.com/nexu-io/open-design) 和 [`multica-ai/multica`](https://github.com/multica-ai/multica) 的 agent 检测策略：唯一被 spawn 子进程的进程是 server route，业务进程不直接 spawn；CLI 的 stdin / stdout 用 JSON-line 协议复用，每个 CLI 一个轻 adapter，统一在 [`next/src/lib/agents/argv.ts`](next/src/lib/agents/argv.ts)。
 

cli/src/__tests__/agents-invoke.test.ts

@@ -527,4 +527,28 @@ describe("invokeAgent", () => {
     });
 
   });
+
+  describe("antigravity", () => {
+    it("passes the prompt right after -p and parses agy step_update/result lines", async () => {
+      const lines = [
+        JSON.stringify({ event: "init", conversation_id: "conv-1", init: { cwd: "/tmp" } }),
+        JSON.stringify({ event: "step_update", step_update: { text_delta: "<p>hi</p>" } }),
+        JSON.stringify({ event: "result", result: { status: "SUCCESS", response: "<p>hi</p>" } }),
+      ].join("\n") + "\n";
+
+      const events = await driveInvoke(
+        { agent: "antigravity", prompt: "make html", binOverride: BIN_OVERRIDE },
+        lines,
+        0,
+      );
+
+      const argv = mockSpawn.mock.calls.at(-1)![1] as string[];
+      expect(argv.slice(-2)).toEqual(["-p", "make html"]);
+      expect(argv).toContain("--disable-slash-commands");
+
+      // result.response must not duplicate the streamed delta
+      const deltas = events.filter((e) => e.type === "delta");
+      expect(deltas).toEqual([{ type: "delta", text: "<p>hi</p>" }]);
+    });
+  });
 });

cli/src/agents-detect.ts

@@ -89,6 +89,16 @@ export const AGENTS: AgentDef[] = [
       { id: "gpt-5", label: "gpt-5" },
     ],
   },
+  {
+    id: "antigravity",
+    label: "Antigravity",
+    bin: "agy",
+    envOverride: "ANTIGRAVITY_BIN",
+    vendor: "Google",
+    protocol: "argv",
+    // Model ids come from the user's agy config; no curated list yet.
+    fallbackModels: [DEFAULT_MODEL],
+  },
   {
     id: "gemini",
     label: "Gemini CLI",

cli/src/agents-invoke.ts

@@ -120,6 +120,20 @@ function buildArgv(agent: string, opts: AgentArgvOpts = {}): string[] {
         "--trust",
         ...(model ? ["--model", model] : []),
       ];
+    case "antigravity":
+      // `-p` must stay last: invoke appends the prompt right after it
+      // (`protocol: "argv"`). Effort is pinned so runs don't depend on the
+      // user's local agy default.
+      return [
+        "--output-format",
+        "stream-json",
+        "--dangerously-skip-permissions",
+        "--disable-slash-commands",
+        "--effort",
+        "medium",
+        ...(model ? ["--model", model] : []),
+        "-p",
+      ];
     case "gemini":
       return [
         "--output-format",
@@ -209,6 +223,7 @@ function rescueHtmlFromToolUse(
       name !== "createfile" &&
       name !== "writefile" &&
       name !== "write_file" &&
+      name !== "write_to_file" &&
       name !== "filewrite"
     )
       continue;
@@ -229,6 +244,20 @@ function rescueHtmlFromToolUse(
   return parts.join("");
 }
 
+/** agy reports `cache_read_tokens`; the UI reads Claude-style usage keys. */
+function agyUsage(u: Record<string, number>): AgentParse {
+  return {
+    kind: "meta",
+    key: "usage",
+    value: {
+      input_tokens: u.input_tokens,
+      output_tokens: u.output_tokens,
+      cache_read_input_tokens: u.cache_read_tokens ?? 0,
+      cache_creation_input_tokens: 0,
+    },
+  };
+}
+
 function parseLineWithState(agent: string, line: string, state: ParseState): AgentParse[] {
   const trimmed = line.trim();
   if (!trimmed) return [];
@@ -246,6 +275,73 @@ function parseLineWithState(agent: string, line: string, state: ParseState): Age
   if (!parsed || typeof parsed !== "object") return [];
   const obj = parsed as Record<string, unknown>;
   const out: AgentParse[] = [];
+  if (agent === "antigravity") {
+    // agy stream-json: one envelope per line, keyed by `event`
+    // (init / step_update / result).
+    if (obj.event === "init") {
+      if (typeof obj.conversation_id === "string") {
+        out.push({ kind: "meta", key: "session", value: obj.conversation_id });
+      }
+      const initObj = obj.init as Record<string, unknown> | undefined;
+      if (typeof initObj?.cwd === "string") {
+        out.push({ kind: "meta", key: "cwd", value: initObj.cwd });
+      }
+    }
+
+    const step = obj.step_update as Record<string, unknown> | undefined;
+    if (step && typeof step === "object") {
+      // agy writes files via `write_to_file { TargetFile, CodeContent }`.
+      // Map it onto the shared tool_use shape so the rescue rules live in
+      // one place.
+      const toolInfo = step.tool_info as
+        | { name?: string; parameters?: Record<string, unknown> }
+        | undefined;
+      const params = toolInfo?.parameters;
+      if (params && (toolInfo?.name ?? step.tool_name) === "write_to_file") {
+        const toolHtml = rescueHtmlFromToolUse([
+          {
+            type: "tool_use",
+            name: "write_to_file",
+            input: { file_path: params.TargetFile, content: params.CodeContent },
+          },
+        ]);
+        if (toolHtml) {
+          out.push({ kind: "html", text: toolHtml });
+          state.sawStreamEventText = true;
+        }
+      }
+      if (typeof step.text_delta === "string" && step.text_delta) {
+        state.sawStreamEventText = true;
+        out.push({ kind: "delta", text: step.text_delta });
+      }
+      if (typeof step.thinking_delta === "string" && step.thinking_delta) {
+        out.push({ kind: "meta", key: "thinking", value: step.thinking_delta });
+      }
+      if (step.usage && typeof step.usage === "object") {
+        out.push(agyUsage(step.usage as Record<string, number>));
+      }
+    }
+
+    const result = obj.result as Record<string, unknown> | undefined;
+    if (result && typeof result === "object") {
+      if (!state.sawStreamEventText && typeof result.response === "string" && result.response) {
+        out.push({ kind: "delta", text: result.response });
+      }
+      if (result.usage && typeof result.usage === "object") {
+        out.push(agyUsage(result.usage as Record<string, number>));
+      }
+      if (typeof result.duration_seconds === "number") {
+        out.push({
+          kind: "meta",
+          key: "duration_ms",
+          value: Math.round(result.duration_seconds * 1000),
+        });
+      }
+      if (typeof result.status === "string") {
+        out.push({ kind: "meta", key: "result", value: result.status });
+      }
+    }
+  }
 
   if (agent === "claude") {
     if (obj.type === "system" && obj.subtype === "init") {