FEAT: 웹 문서 가져오기 도구

app/e2e/run.mjs

@@ -74,11 +74,17 @@ const file = (rel) => fs.readFileSync(path.join(DIRS.proj, rel), "utf8");
 
 // ── 모의 모델 (OpenAI 호환, 스트리밍) ────────────────────────
 // 1차: read_file → 2차: edit_file → 3차: 완료. 무엇을 읽고 고칠지는 시나리오가 정한다.
-const mock = { read: "src/auth/login.ts", edit: "src/auth/session.ts", old: "v + '.sig'", new: "v + '.signed'", requests: 0, testPrompt: "", image: null, embedded: 0, semanticPrompt: "", toolNames: [], verifyFailure: "" };
+const mock = { read: "src/auth/login.ts", edit: "src/auth/session.ts", old: "v + '.sig'", new: "v + '.signed'", requests: 0, testPrompt: "", image: null, embedded: 0, semanticPrompt: "", toolNames: [], verifyFailure: "", fetched: [], fetchResult: "" };
 const mockServer = http.createServer((req, res) => {
   let body = "";
   req.on("data", (c) => (body += c));
   req.on("end", () => {
+    // fetch_url이 읽을 가짜 문서
+    if (req.method === "GET" && req.url.startsWith("/docs/")) {
+      mock.fetched.push(req.url);
+      res.writeHead(200, { "content-type": "text/html; charset=utf-8" });
+      return res.end(`<html><head><script>steal()</script></head><body><h1>${req.url === "/docs/cookie" ? "Cookie signing" : "Session lifetime"}</h1><p>Sign cookies with an HMAC; keep sessions for 7 days.</p></body></html>`);
+    }
     if (req.method === "GET") {
       res.writeHead(200, { "content-type": "application/json" });
       return res.end(JSON.stringify({ data: [{ id: "e2e-model" }] }));
@@ -143,6 +149,19 @@ const mockServer = http.createServer((req, res) => {
       }
       return edit("v + '.sig'", "v + '.bad'");
     }
+    // 웹 문서: 같은 사이트의 두 페이지를 읽고 요약한다
+    if (first.includes("문서를 찾아봐")) {
+      const fetchCall = (url) => {
+        send({ choices: [{ index: 0, delta: { tool_calls: [{ index: 0, id: `call_f${tools}`, type: "function", function: { name: "fetch_url", arguments: JSON.stringify({ url }) } }] } }] });
+        done("tool_calls");
+      };
+      if (tools === 0) return fetchCall(`http://127.0.0.1:${MOCK_PORT}/docs/cookie`);
+      if (tools === 1) return fetchCall(`http://127.0.0.1:${MOCK_PORT}/docs/session`);
+      const result = String(j.messages.filter((m) => m.role === "tool").at(0)?.content ?? "");
+      mock.fetchResult = result;
+      send({ choices: [{ index: 0, delta: { content: "문서 요약: HMAC으로 서명" } }] });
+      return done("stop");
+    }
     // 외부 MCP 도구: 티켓을 찾아 그 내용으로 답한다
     if (first.includes("티켓 T-1")) {
       mock.toolNames = (j.tools ?? []).map((t) => t.function?.name);
@@ -233,6 +252,7 @@ async function launch(dir = DIRS.proj) {
       LANTERN_WEBVIEW_DATA_DIR: DIRS.wv,
       LANTERN_INDEX_DIR: DIRS.index,
       LANTERN_E2E_CDP_PORT: String(cdpPort),
+      LANTERN_FETCH_ALLOW_LOCAL: "1", // 가짜 문서 서버가 로컬이라
     },
     stdio: "ignore",
   });
@@ -915,6 +935,29 @@ scenario("고친 뒤 관련 테스트로 확인: 실패 → 출력을 받아 다
   git("checkout", "--", "src/auth/session.ts");
 });
 
+scenario("웹 문서 가져오기: 처음 여는 사이트만 승인 → 본문 글로 읽기", async () => {
+  await run(() => {
+    document.querySelector("#btn-new-task").click();
+    return true;
+  });
+  await run(ui.sendTask, "쿠키 서명 문서를 찾아봐", "ask", false);
+  const card = await waitFor(() => {
+    const c = document.querySelector(".task-log:not(.hidden) .approval:not(.resolved)");
+    return c && { title: c.querySelector(".atitle")?.textContent ?? "", globe: !!c.querySelector(".ahead .codicon-globe"), detail: c.querySelector("pre")?.textContent ?? "" };
+  }, 20000, "사이트 승인 카드");
+  if (!card.title.includes("127.0.0.1") || !card.globe || !card.detail.includes("/docs/cookie")) throw new Error(`승인 카드: ${JSON.stringify(card)}`);
+  if (mock.fetched.length) throw new Error("승인 전에 가져옴");
+  await run(() => {
+    [...document.querySelectorAll(".task-log:not(.hidden) .approval:not(.resolved) .actions button")].find((b) => b.textContent.includes("가져오기")).click();
+    return true;
+  });
+  await waitFor(() => document.querySelector(".task-log:not(.hidden)")?.innerText.includes("문서 요약"), 20000, "요약");
+  // 같은 사이트의 두 번째 페이지는 묻지 않는다
+  const approvals = await run(() => document.querySelectorAll(".task-log:not(.hidden) .approval").length);
+  if (approvals !== 1 || mock.fetched.join() !== "/docs/cookie,/docs/session") throw new Error(`승인 ${approvals}개, 가져온 곳 ${mock.fetched}`);
+  if (!mock.fetchResult.includes("Cookie signing") || !mock.fetchResult.includes("HMAC") || mock.fetchResult.includes("steal()") || mock.fetchResult.includes("<p>")) throw new Error(`본문: ${mock.fetchResult}`);
+});
+
 scenario("편집기 안 즉시 수정 (Ctrl+K): diff·영향 반경 → 적용 → 되돌리기", async () => {
   await run(() => {
     document.querySelector('.ab-item[data-view="explorer"]').click();

app/src-tauri/Cargo.toml

@@ -17,7 +17,7 @@ anyhow = "1"
 serde = { version = "1", features = ["derive"] }
 serde_json = "1"
 toml = "1.1"
-tokio = { version = "1", features = ["rt-multi-thread", "macros", "sync", "time", "process", "io-util"] }
+tokio = { version = "1", features = ["rt-multi-thread", "macros", "sync", "time", "process", "io-util", "net"] }
 reqwest = { version = "0.12", default-features = false, features = ["json", "stream", "rustls-tls"] }
 futures-util = "0.3"
 portable-pty = "0.9"
@@ -32,6 +32,7 @@ notify = "8"
 trash = "5"
 tauri-plugin-updater = "2"
 regex = "1"
+html2text = "0.17.1"
 
 [dev-dependencies]
 tempfile = "3"

app/src-tauri/src/agent.rs

@@ -352,7 +352,7 @@ async fn run_inner(
     push_user(&mut history, format!("{text}\n\n<project_context>\n{}\n</project_context>", ctx.0), images);
 
     let approver = TauriApprover { app: app.clone(), session: session.to_string(), cancel: cancel.clone() };
-    let tctx = ToolCtx { project: project.clone(), config: &config, approver: &approver, changed: Default::default(), originals: Default::default(), workdir: workdir.clone(), mcp: mcp.map };
+    let tctx = ToolCtx { project: project.clone(), config: &config, approver: &approver, changed: Default::default(), originals: Default::default(), workdir: workdir.clone(), mcp: mcp.map, fetched_hosts: Default::default() };
     let max_steps = config.agent.max_steps.max(1);
     // 고친 뒤 확인: 쓰기·실행이 되는 에이전트, 신뢰한 폴더에서만
     let verify_on = config.agent.verify_tests && project.is_trusted() && allowed.iter().any(|t| t == "run_command") && allowed.iter().any(|t| t == "edit_file" || t == "write_file");

app/src-tauri/src/agents.rs

@@ -14,14 +14,15 @@ use serde::Serialize;
 use std::path::Path;
 
 pub const READ_TOOLS: &[&str] =
-    &["get_context", "search_symbols", "get_symbol", "find_references", "read_file", "list_dir"];
+    &["get_context", "search_symbols", "get_symbol", "find_references", "read_file", "list_dir", "fetch_url"];
 pub const ALL_TOOLS: &[&str] = &[
     "get_context",
     "search_symbols",
     "get_symbol",
     "find_references",
     "read_file",
     "list_dir",
+    "fetch_url",
     "write_file",
     "edit_file",
     "run_command",

app/src-tauri/src/config.rs

@@ -206,6 +206,9 @@ pub struct AgentConfig {
     pub auto_approve: Vec<String>,
     #[serde(default)]
     pub allowed_commands: Vec<String>,
+    /// fetch_url이 승인 없이 열 사이트 (도메인과 그 하위 도메인)
+    #[serde(default)]
+    pub allowed_domains: Vec<String>,
     /// 파일을 고치고 끝낼 때 관련 테스트를 돌려 확인한다 (실패하면 다시 고치게)
     #[serde(default = "default_true")]
     pub verify_tests: bool,

app/src-tauri/src/e2e_tests.rs

@@ -234,7 +234,7 @@ async fn tools_edit_with_approval_and_rejection() {
 
     // 승인하는 경우
     let yes = FakeApprover { answer: true, asked: Mutex::new(vec![]) };
-    let ctx = ToolCtx { project: p.clone(), config: &cfg, approver: &yes, changed: Default::default(), originals: Default::default(), workdir: None, mcp: Default::default() };
+    let ctx = ToolCtx { project: p.clone(), config: &cfg, approver: &yes, changed: Default::default(), originals: Default::default(), workdir: None, mcp: Default::default(), fetched_hosts: Default::default() };
     let out = ctx
         .run("edit_file", &json!({"path":"src/lib.rs","old_string":"hi {name}","new_string":"안녕 {name}"}))
         .await;
@@ -259,7 +259,7 @@ async fn tools_edit_with_approval_and_rejection() {
 
     // 거절하는 경우: 파일이 그대로여야 한다
     let no = FakeApprover { answer: false, asked: Mutex::new(vec![]) };
-    let ctx2 = ToolCtx { project: p.clone(), config: &cfg, approver: &no, changed: Default::default(), originals: Default::default(), workdir: None, mcp: Default::default() };
+    let ctx2 = ToolCtx { project: p.clone(), config: &cfg, approver: &no, changed: Default::default(), originals: Default::default(), workdir: None, mcp: Default::default(), fetched_hosts: Default::default() };
     let rejected = ctx2.run("write_file", &json!({"path":"src/new.rs","content":"x"})).await;
     assert!(rejected.is_error && rejected.content.contains("거절"));
     assert!(!dir.path().join("src/new.rs").exists());
@@ -300,7 +300,7 @@ async fn tools_edit_with_approval_and_rejection() {
     // 자동 승인 glob
     cfg.agent.auto_approve = vec!["docs/**".into()];
     let never = FakeApprover { answer: false, asked: Mutex::new(vec![]) };
-    let ctx3 = ToolCtx { project: p.clone(), config: &cfg, approver: &never, changed: Default::default(), originals: Default::default(), workdir: None, mcp: Default::default() };
+    let ctx3 = ToolCtx { project: p.clone(), config: &cfg, approver: &never, changed: Default::default(), originals: Default::default(), workdir: None, mcp: Default::default(), fetched_hosts: Default::default() };
     let auto = ctx3.run("write_file", &json!({"path":"docs/a.md","content":"# a"})).await;
     assert!(!auto.is_error, "{}", auto.content);
     assert!(never.asked.lock().unwrap().is_empty());

app/src-tauri/src/main.rs

@@ -11,6 +11,7 @@ mod review;
 mod semantic;
 mod mcp_client;
 mod verify;
+mod webfetch;
 mod tasks;
 mod worktree;
 mod completion;

app/src-tauri/src/tools.rs

@@ -31,6 +31,8 @@ pub fn specs(allowed: &[String]) -> Vec<ToolSpec> {
             json!({"type":"object","properties":{"name":{"type":"string"}},"required":["name"]})),
         ("read_file", "Read a file with line numbers (`N<TAB>text`). Optionally a 1-based inclusive line range.",
             json!({"type":"object","properties":{"path":{"type":"string"},"start_line":{"type":"integer"},"end_line":{"type":"integer"}},"required":["path"]})),
+        ("fetch_url", "Fetch a web page (http/https) and return it as text, e.g. a library's documentation or an explanation of an error message. Each new site needs the user's approval. Never put project code, file contents or secrets in the URL.",
+            json!({"type":"object","properties":{"url":{"type":"string"}},"required":["url"]})),
         ("list_dir", "List a directory (respects .gitignore). Use \".\" for the project root.",
             json!({"type":"object","properties":{"path":{"type":"string"}},"required":["path"]})),
         ("write_file", "Create or overwrite a file with the full content. The user reviews a diff before it is applied. Prefer edit_file for small changes to existing files.",
@@ -70,6 +72,8 @@ pub struct ToolCtx<'a> {
     pub workdir: Option<std::path::PathBuf>,
     /// 외부 MCP 도구: 모델에 보이는 이름 → (서버 연결, 서버의 도구 이름)
     pub mcp: HashMap<String, (Arc<crate::mcp_client::Conn>, String)>,
+    /// 이번 작업에서 사용자가 열어도 된다고 한 사이트
+    pub fetched_hosts: std::sync::Mutex<std::collections::HashSet<String>>,
 }
 
 pub struct ToolOutcome {
@@ -254,6 +258,17 @@ impl ToolCtx<'_> {
                 self.approver.output(&out);
                 Ok(out)
             }
+            "fetch_url" => {
+                let url = crate::webfetch::check_url(s(input, "url")?)?;
+                let host = url.host_str().unwrap_or_default().to_ascii_lowercase();
+                let known = crate::webfetch::host_allowed(&host, &self.config.agent.allowed_domains) || self.fetched_hosts.lock().unwrap().contains(&host);
+                if !known && !self.approver.ask("fetch", &format!("웹 페이지 가져오기: {host}"), url.as_str()).await {
+                    bail!("사용자가 {host} 접속을 거절했습니다");
+                }
+                self.fetched_hosts.lock().unwrap().insert(host);
+                let f = crate::webfetch::fetch(url.as_str()).await?;
+                Ok(format!("{}\n\n{}", f.url, f.text.trim()))
+            }
             other if other.starts_with(crate::mcp_client::PREFIX) => self.mcp_call(other, input).await,
             other => bail!("알 수 없는 도구: {other}"),
         }