assistme 0.3.1 → 0.3.3

package/src/tools/index.ts

@@ -1,4 +1,10 @@
-import { type BrowserController, getBrowser, ensureBrowserAvailable } from "./browser.js";
+import {
+  BrowserController,
+  getBrowser,
+  ensureBrowserAvailable,
+  type SnapshotResult,
+  type ActionSpec,
+} from "./browser.js";
 import {
   readFileContent,
   writeFileContent,
@@ -8,268 +14,6 @@ import {
 } from "./filesystem.js";
 import { executeShell } from "./shell.js";
 
-interface ToolDefinition {
-  name: string;
-  description: string;
-  input_schema: {
-    type: "object";
-    properties: Record<string, unknown>;
-    required?: string[];
-  };
-}
-
-export function getToolDefinitions(): ToolDefinition[] {
-  return [
-    // ── File System Tools ───────────────────────────────────────
-    {
-      name: "read_file",
-      description:
-        "Read the contents of a file. Returns line-numbered content. Use offset/limit for large files.",
-      input_schema: {
-        type: "object",
-        properties: {
-          path: { type: "string", description: "File path relative to workspace" },
-          offset: { type: "number", description: "Start line (0-indexed)" },
-          limit: { type: "number", description: "Max lines to read" },
-        },
-        required: ["path"],
-      },
-    },
-    {
-      name: "write_file",
-      description: "Write content to a file. Creates parent directories if needed.",
-      input_schema: {
-        type: "object",
-        properties: {
-          path: { type: "string", description: "File path relative to workspace" },
-          content: { type: "string", description: "Content to write" },
-        },
-        required: ["path", "content"],
-      },
-    },
-    {
-      name: "search_files",
-      description: "Search for files matching a glob pattern (e.g. '**/*.ts').",
-      input_schema: {
-        type: "object",
-        properties: {
-          pattern: { type: "string", description: "Glob pattern" },
-          directory: { type: "string", description: "Directory (relative to workspace)" },
-        },
-        required: ["pattern"],
-      },
-    },
-    {
-      name: "search_content",
-      description: "Search file contents for a regex pattern. Returns matching lines.",
-      input_schema: {
-        type: "object",
-        properties: {
-          pattern: { type: "string", description: "Regex pattern" },
-          file_glob: { type: "string", description: "File glob filter (default: **/*)" },
-          directory: { type: "string", description: "Directory (relative to workspace)" },
-        },
-        required: ["pattern"],
-      },
-    },
-    {
-      name: "list_directory",
-      description: "List files and directories in a path.",
-      input_schema: {
-        type: "object",
-        properties: {
-          path: { type: "string", description: "Directory path (default: workspace root)" },
-        },
-      },
-    },
-    {
-      name: "execute_command",
-      description: "Execute a shell command in the workspace directory.",
-      input_schema: {
-        type: "object",
-        properties: {
-          command: { type: "string", description: "Shell command" },
-          cwd: { type: "string", description: "Working directory (relative)" },
-        },
-        required: ["command"],
-      },
-    },
-
-    // ── Browser Tools (CDP - controls user's real Chrome) ───────
-    {
-      name: "browser_connect",
-      description:
-        "Connect to the user's real Chrome browser via CDP. Chrome will be auto-launched if not already running. This shares the user's actual browser session including all logins and cookies.",
-      input_schema: {
-        type: "object",
-        properties: {
-          tab_index: {
-            type: "number",
-            description:
-              "Tab index to connect to (default: 0, the first tab). Use browser_list_tabs to see available tabs.",
-          },
-        },
-      },
-    },
-    {
-      name: "browser_navigate",
-      description:
-        "Navigate the user's browser to a URL. Opens the page in the currently connected tab, using the user's real browser with all their cookies and logins.",
-      input_schema: {
-        type: "object",
-        properties: {
-          url: { type: "string", description: "URL to navigate to" },
-        },
-        required: ["url"],
-      },
-    },
-    {
-      name: "browser_read_page",
-      description:
-        "Read the text content of the currently open page in the user's browser. Returns the page title, URL, and main text content. Use this to understand what's on the page.",
-      input_schema: { type: "object", properties: {} },
-    },
-    {
-      name: "browser_screenshot",
-      description:
-        "Take a screenshot of the current browser page. Returns a base64-encoded PNG image. Use this when you need to visually understand the page layout, verify an action worked, or see dynamic content.",
-      input_schema: { type: "object", properties: {} },
-    },
-    {
-      name: "browser_click",
-      description:
-        "Click on an element in the user's browser using a CSS selector. Use browser_get_elements first to find clickable elements.",
-      input_schema: {
-        type: "object",
-        properties: {
-          selector: {
-            type: "string",
-            description:
-              "CSS selector of the element to click (e.g. '#submit-btn', 'a.nav-link', 'button:nth-of-type(2)')",
-          },
-        },
-        required: ["selector"],
-      },
-    },
-    {
-      name: "browser_type",
-      description:
-        "Type text into an input field in the user's browser. Focuses the element and sets its value.",
-      input_schema: {
-        type: "object",
-        properties: {
-          selector: {
-            type: "string",
-            description: "CSS selector of the input element",
-          },
-          text: {
-            type: "string",
-            description: "Text to type into the element",
-          },
-        },
-        required: ["selector", "text"],
-      },
-    },
-    {
-      name: "browser_press_key",
-      description:
-        "Press a keyboard key in the browser. Supports: Enter, Tab, Escape, Backspace, ArrowDown, ArrowUp, or any single character.",
-      input_schema: {
-        type: "object",
-        properties: {
-          key: { type: "string", description: "Key to press" },
-        },
-        required: ["key"],
-      },
-    },
-    {
-      name: "browser_scroll",
-      description: "Scroll the page up or down.",
-      input_schema: {
-        type: "object",
-        properties: {
-          direction: {
-            type: "string",
-            description: "'down' or 'up'",
-          },
-        },
-        required: ["direction"],
-      },
-    },
-    {
-      name: "browser_get_elements",
-      description:
-        "Find all interactive elements (links, buttons, inputs, etc.) on the current page. Returns their tag, text, selector, and attributes. Use this to understand what you can interact with.",
-      input_schema: { type: "object", properties: {} },
-    },
-    {
-      name: "browser_evaluate",
-      description:
-        "Execute JavaScript in the browser page context. Use for advanced interactions or data extraction that other tools can't handle.",
-      input_schema: {
-        type: "object",
-        properties: {
-          expression: {
-            type: "string",
-            description: "JavaScript expression to evaluate",
-          },
-        },
-        required: ["expression"],
-      },
-    },
-    {
-      name: "browser_list_tabs",
-      description:
-        "List all open tabs in the user's browser. Shows title, URL, and which tab is currently connected.",
-      input_schema: { type: "object", properties: {} },
-    },
-    {
-      name: "browser_switch_tab",
-      description: "Switch to a different browser tab by index.",
-      input_schema: {
-        type: "object",
-        properties: {
-          index: {
-            type: "number",
-            description: "Tab index (from browser_list_tabs)",
-          },
-        },
-        required: ["index"],
-      },
-    },
-    {
-      name: "browser_new_tab",
-      description: "Open a new tab in the user's browser, optionally navigating to a URL.",
-      input_schema: {
-        type: "object",
-        properties: {
-          url: { type: "string", description: "URL to open (default: blank)" },
-        },
-      },
-    },
-    {
-      name: "browser_request_user_action",
-      description:
-        "Request the user to perform an action in their browser that the AI cannot do. Use this when: (1) a page requires login/authentication, (2) a CAPTCHA needs solving, (3) two-factor auth is needed, (4) any action requiring the user's personal credentials. The user will see a notification on the web UI and in the terminal.",
-      input_schema: {
-        type: "object",
-        properties: {
-          message: {
-            type: "string",
-            description:
-              "Clear description of what the user needs to do. E.g.: 'Please log in to amazon.com in your browser. The login page is currently open in the active tab.'",
-          },
-          wait_seconds: {
-            type: "number",
-            description: "How long to wait for the user to complete the action (default: 60)",
-          },
-        },
-        required: ["message"],
-      },
-    },
-  ];
-}
-
 /**
  * After navigation, detect if the page requires login and automatically prompt
  * the user to log in. Waits for login completion and returns context about the
@@ -425,12 +169,34 @@ export async function executeTool(name: string, input: Record<string, unknown>):
     case "browser_scroll":
       await ensureConnected(browser);
       return (input.direction as string) === "up" ? browser.scrollUp() : browser.scrollDown();
-    case "browser_get_elements":
+    case "browser_select":
       await ensureConnected(browser);
-      return browser.getInteractiveElements();
+      return browser.selectOption(input.selector as string, input.option as string);
     case "browser_evaluate":
       await ensureConnected(browser);
       return browser.evaluate(input.expression as string);
+    case "browser_snapshot": {
+      await ensureConnected(browser);
+      const snap = await browser.snapshot(input.annotate as boolean | undefined);
+      // Return is handled specially in MCP server (image + text)
+      // For the generic executeTool path, return the ref table text
+      return BrowserController.formatRefTable(snap) + "\n__SNAPSHOT_IMAGE__:" + snap.image;
+    }
+    case "browser_act": {
+      await ensureConnected(browser);
+      const actions = input.actions as ActionSpec[];
+      const wantScreenshot = (input.screenshot as boolean) || false;
+      const actResult = await browser.act(actions, wantScreenshot);
+
+      let response = actResult.results
+        .map((r) => `${r.success ? "OK" : "FAIL"}: ${r.result}`)
+        .join("\n");
+
+      if (actResult.screenshot) {
+        response += "\n__ACT_SCREENSHOT__:" + actResult.screenshot;
+      }
+      return response;
+    }
     case "browser_list_tabs":
       return browser.listTabs();
     case "browser_switch_tab":

package/src/tools/web.ts

@@ -1,73 +0,0 @@
-const MAX_CONTENT = 30_000; // 30KB content cap
-
-export async function webFetch(
-  url: string,
-  _prompt?: string
-): Promise<string> {
-  try {
-    const controller = new AbortController();
-    const timeoutId = setTimeout(() => controller.abort(), 15_000);
-
-    const response = await fetch(url, {
-      signal: controller.signal,
-      headers: {
-        "User-Agent": "AssistMe-CLI/0.1.0",
-        Accept: "text/html,application/json,text/plain",
-      },
-    });
-
-    clearTimeout(timeoutId);
-
-    if (!response.ok) {
-      return `HTTP ${response.status}: ${response.statusText}`;
-    }
-
-    const contentType = response.headers.get("content-type") || "";
-    let content: string;
-
-    if (contentType.includes("application/json")) {
-      const json = await response.json();
-      content = JSON.stringify(json, null, 2);
-    } else {
-      content = await response.text();
-
-      // Basic HTML to text conversion
-      if (contentType.includes("text/html")) {
-        content = htmlToText(content);
-      }
-    }
-
-    if (content.length > MAX_CONTENT) {
-      content =
-        content.slice(0, MAX_CONTENT) + "\n\n[Content truncated]";
-    }
-
-    return content;
-  } catch (err) {
-    return `Fetch error: ${err instanceof Error ? err.message : String(err)}`;
-  }
-}
-
-function htmlToText(html: string): string {
-  return html
-    // Remove scripts and styles
-    .replace(/<script[^>]*>[\s\S]*?<\/script>/gi, "")
-    .replace(/<style[^>]*>[\s\S]*?<\/style>/gi, "")
-    // Convert common elements
-    .replace(/<br\s*\/?>/gi, "\n")
-    .replace(/<\/p>/gi, "\n\n")
-    .replace(/<\/div>/gi, "\n")
-    .replace(/<\/h[1-6]>/gi, "\n\n")
-    .replace(/<li>/gi, "- ")
-    .replace(/<\/li>/gi, "\n")
-    // Remove remaining tags
-    .replace(/<[^>]+>/g, "")
-    // Clean up whitespace
-    .replace(/&nbsp;/g, " ")
-    .replace(/&amp;/g, "&")
-    .replace(/&lt;/g, "<")
-    .replace(/&gt;/g, ">")
-    .replace(/&quot;/g, '"')
-    .replace(/\n{3,}/g, "\n\n")
-    .trim();
-}