agentic-browser 2.6.1 → 2.6.2

package/AGENTS.md

@@ -3,14 +3,14 @@
 ## Quick Reference
 
 ```bash
-npm run build        # tsdown (~20ms)
-npm run typecheck    # tsc --noEmit
-npm run lint         # oxlint
-npm run lint:fix     # oxlint --fix
-npm run format       # oxfmt --write
-npm test             # vitest run
-npm run test:watch   # vitest
-npm run docs:dev     # vocs dev server
+pnpm build # tsdown (~20ms)
+pnpm typecheck # tsc --noEmit
+pnpm lint # oxlint
+pnpm lint:fix # oxlint --fix
+pnpm format # oxfmt --write
+pnpm test # vitest run
+pnpm test:watch # vitest
+pnpm docs:dev # vocs dev server
 ```
 
 ## Architecture
@@ -29,7 +29,7 @@ src/
     commands/agent.ts     — Stateful agent commands (auto-restart, retry, session persistence)
     commands/*.ts         — Low-level CLI command handlers
   mcp/
-    index.ts              — MCP server (stdio transport, 7 tools wrapping AgenticBrowserCore)
+    index.ts              — MCP server (stdio transport, tools wrapping AgenticBrowserCore)
   session/
     browser-controller.ts — BrowserController interface + ChromeCdpBrowserController (CDP WebSocket)
     session-manager.ts    — Orchestrates sessions, commands, memory recording
@@ -68,7 +68,7 @@ AgenticBrowserCore → ControlApi → SessionManager → BrowserController (CDP)
 
 - **ESM-only**: `"type": "module"`, use `.js` extensions in all TypeScript imports
 - **Zod v4**: `import { z } from "zod"` — `z.record()` requires key+value args
-- **Commander.js v14**: colon-namespaced commands (`session:start`, `memory:search`)
+- **Commander.js v15**: colon-namespaced commands (`session:start`, `memory:search`)
 - **CLI output**: exactly one JSON line to `stdout`, errors to `stderr`
 - **Types**: interfaces for public contracts, type aliases for unions/inferred
 - **No console.log in MCP server**: use `process.stderr.write()` for debug output
@@ -111,21 +111,27 @@ AgenticBrowserCore → ControlApi → SessionManager → BrowserController (CDP)
 - `AGENTIC_BROWSER_DIR` — base dir for sessions/memory/events (default: `~/.agentic-browser`)
 - `AGENTIC_BROWSER_CHROME_EXECUTABLE_PATH` — explicit Chrome path (auto-discovered if not set)
 - `AGENTIC_BROWSER_CDP_URL` — connect to an already-running Chrome via CDP (e.g. `http://127.0.0.1:9222`)
+- `AGENTIC_BROWSER_ALLOW_FILE_URLS` — allow `file:` URLs in navigate (`true` or `1`); default is http(s) and about:blank only
 - `AGENTIC_BROWSER_USER_PROFILE` — use real Chrome profile (`default`, `true`, or absolute path)
+- `AGENTIC_BROWSER_CHROME_PATH` — alias of `AGENTIC_BROWSER_CHROME_EXECUTABLE_PATH`
+- `AGENTIC_BROWSER_HEADLESS` — `true` or `1` to launch Chrome headless
+- `AGENTIC_BROWSER_USER_AGENT` — override the browser user agent
 
 ## MCP Server
 
 Subcommand: `agentic-browser mcp` (stdio transport). Setup: `agentic-browser setup`. Tools:
 
-| Tool                    | Purpose                                                                                                   |
-| ----------------------- | --------------------------------------------------------------------------------------------------------- |
-| `browser_start_session` | Start Chrome, return sessionId                                                                            |
-| `browser_navigate`      | Navigate to URL                                                                                           |
-| `browser_interact`      | click / type / press / waitFor / scroll / hover / select / toggle / goBack / goForward / refresh / dialog |
-| `browser_get_content`   | Get page title / text / html                                                                              |
-| `browser_get_elements`  | Discover interactive elements                                                                             |
-| `browser_search_memory` | Search task memory                                                                                        |
-| `browser_stop_session`  | Stop Chrome session                                                                                       |
+| Tool                      | Purpose                                                                                                   |
+| ------------------------- | --------------------------------------------------------------------------------------------------------- |
+| `browser_start_session`   | Start Chrome, return sessionId                                                                            |
+| `browser_navigate`        | Navigate to URL                                                                                           |
+| `browser_interact`        | click / type / press / waitFor / scroll / hover / select / toggle / goBack / goForward / refresh / dialog |
+| `browser_get_content`     | Get page title / text / html                                                                              |
+| `browser_get_elements`    | Discover interactive elements                                                                             |
+| `browser_search_memory`   | Search task memory                                                                                        |
+| `browser_stop_session`    | Stop Chrome session                                                                                       |
+| `browser_dismiss_cookies` | Dismiss cookie or consent banner                                                                          |
+| `browser_screenshot`      | Screenshot as base64 (jpeg or png, optional full page)                                                    |
 
 ## For Browser Automation Tasks
 

package/README.md

@@ -13,7 +13,7 @@ CLI and MCP server to control a local Chrome session for AI agents.
 
 ## Requirements
 
-- Node.js 20+
+- Node.js 22+
 - Installed Chrome
 
 ## Install

package/dist/cli/index.mjs

@@ -1,10 +1,10 @@
 #!/usr/bin/env node
-import { r as createCliRuntime } from "../runtime-BzdbtFF4.mjs";
+import { a as isConnectionError, o as ensurePrivateDir, r as createCliRuntime, s as writePrivateFileSync } from "../runtime-Bo7FmRYk.mjs";
 import fs from "node:fs";
 import path from "node:path";
 import crypto from "node:crypto";
 import { Command } from "commander";
-//#region src/cli/commands/agent.ts
+// src/cli/commands/agent.ts
 const STATE_FILE_NAME = "agent-state.json";
 function stateFilePath(runtime) {
 	return path.join(runtime.context.config.dataDir, STATE_FILE_NAME);
@@ -20,8 +20,8 @@ function loadState(runtime) {
 }
 function saveState(runtime, state) {
 	const dir = path.dirname(stateFilePath(runtime));
-	fs.mkdirSync(dir, { recursive: true });
-	fs.writeFileSync(stateFilePath(runtime), JSON.stringify(state, null, 2));
+	ensurePrivateDir(dir);
+	writePrivateFileSync(stateFilePath(runtime), JSON.stringify(state, null, 2));
 }
 function requireSessionId(runtime) {
 	const state = loadState(runtime);
@@ -63,6 +63,12 @@ async function agentStop(runtime) {
 		sessionId
 	};
 }
+/** Only connection-level failures justify a restart + re-run. */
+function isRetryableFailure(message) {
+	if (isConnectionError(message)) return true;
+	const lower = message.toLowerCase();
+	return lower.includes("session") && (lower.includes("dead") || lower.includes("terminated"));
+}
 async function executeAgentCommand(runtime, input) {
 	const sessionId = requireSessionId(runtime);
 	await ensureReady(runtime, sessionId);
@@ -71,16 +77,24 @@ async function executeAgentCommand(runtime, input) {
 		type: input.type,
 		payload: input.payload
 	});
-	try {
-		const first = await attempt();
-		if (first.resultStatus === "failed") throw new Error(first.resultMessage ?? "Command failed");
-		return first;
-	} catch {
+	const retryAfterRestart = async () => {
 		await runtime.api.restartSession(sessionId);
 		const second = await attempt();
 		if (second.resultStatus === "failed") throw new Error(second.resultMessage ?? "Command failed after retry");
 		return second;
+	};
+	const noRetry = input.type === "restart" || input.type === "terminate";
+	let first;
+	try {
+		first = await attempt();
+	} catch (error) {
+		if (noRetry || !isRetryableFailure(error.message)) throw error;
+		return await retryAfterRestart();
 	}
+	if (first.resultStatus !== "failed") return first;
+	const message = first.resultMessage ?? "Command failed";
+	if (noRetry || !isRetryableFailure(message)) throw new Error(message);
+	return await retryAfterRestart();
 }
 async function agentNavigate(runtime, input) {
 	const result = await executeAgentCommand(runtime, {
@@ -184,16 +198,14 @@ async function agentCleanup(runtime, input) {
 		...runtime.api.cleanupSessions(input)
 	};
 }
-//#endregion
-//#region src/cli/output/result-formatter.ts
+// src/cli/output/result-formatter.ts
 function formatResult(resultStatus, message) {
 	return {
 		resultStatus,
 		message
 	};
 }
-//#endregion
-//#region src/cli/commands/session-start.ts
+// src/cli/commands/session-start.ts
 async function runSessionStart(runtime, input) {
 	const session = await runtime.api.createSession(input);
 	return {
@@ -201,29 +213,24 @@ async function runSessionStart(runtime, input) {
 		sessionId: session.sessionId
 	};
 }
-//#endregion
-//#region src/cli/commands/session-status.ts
+// src/cli/commands/session-status.ts
 async function runSessionStatus(runtime, input) {
 	return runtime.api.getSession(input.sessionId);
 }
-//#endregion
-//#region src/cli/commands/session-stop.ts
+// src/cli/commands/session-stop.ts
 async function runSessionStop(runtime, input) {
 	await runtime.api.terminateSession(input.sessionId);
 	return { terminated: true };
 }
-//#endregion
-//#region src/cli/commands/session-cleanup.ts
+// src/cli/commands/session-cleanup.ts
 async function runSessionCleanup(runtime, input) {
 	return runtime.api.cleanupSessions(input);
 }
-//#endregion
-//#region src/cli/commands/session-restart.ts
+// src/cli/commands/session-restart.ts
 async function runSessionRestart(runtime, input) {
 	return await runtime.api.restartSession(input.sessionId);
 }
-//#endregion
-//#region src/cli/commands/session-auth.ts
+// src/cli/commands/session-auth.ts
 async function runSessionAuth(runtime, input) {
 	const token = runtime.api.rotateSessionToken(input.sessionId);
 	return {
@@ -231,8 +238,7 @@ async function runSessionAuth(runtime, input) {
 		valid: runtime.context.tokenService.validate(input.sessionId, token)
 	};
 }
-//#endregion
-//#region src/cli/commands/command-run.ts
+// src/cli/commands/command-run.ts
 async function runCommand(runtime, input) {
 	if (!input.sessionId) throw new Error("No active session. Start or restart a session first.");
 	try {
@@ -245,8 +251,7 @@ async function runCommand(runtime, input) {
 		throw new Error(`Command rejected: ${error.message}`, { cause: error });
 	}
 }
-//#endregion
-//#region src/cli/commands/page-content.ts
+// src/cli/commands/page-content.ts
 async function runPageContent(runtime, input) {
 	return await runtime.api.getContent(input.sessionId, {
 		mode: input.mode,
@@ -254,31 +259,26 @@ async function runPageContent(runtime, input) {
 		maxChars: input.maxChars
 	});
 }
-//#endregion
-//#region src/cli/commands/memory-search.ts
+// src/cli/commands/memory-search.ts
 async function runMemorySearch(runtime, input) {
 	if (!input.taskIntent || !input.taskIntent.trim()) throw new Error("taskIntent is required");
 	return runtime.api.searchMemory(input);
 }
-//#endregion
-//#region src/cli/commands/memory-inspect.ts
+// src/cli/commands/memory-inspect.ts
 async function runMemoryInspect(runtime, input) {
 	if (!input.insightId || !input.insightId.trim()) throw new Error("insightId is required");
 	return runtime.api.inspectMemory(input.insightId);
 }
-//#endregion
-//#region src/cli/commands/memory-verify.ts
+// src/cli/commands/memory-verify.ts
 async function runMemoryVerify(runtime, input) {
 	if (!input.insightId || !input.insightId.trim()) throw new Error("insightId is required");
 	return runtime.api.verifyMemory(input.insightId);
 }
-//#endregion
-//#region src/cli/commands/memory-stats.ts
+// src/cli/commands/memory-stats.ts
 async function runMemoryStats(runtime) {
 	return runtime.api.memoryStats();
 }
-//#endregion
-//#region src/cli/index.ts
+// src/cli/index.ts
 const collect = (value, previous = []) => [...previous, value];
 async function main() {
 	const runtime = createCliRuntime();
@@ -513,7 +513,7 @@ async function main() {
 		await startMcpServer();
 	});
 	program.command("setup").description("Configure agentic-browser as MCP server for your AI tool").action(async () => {
-		const { runSetup } = await import("../setup-DUAjdbQ3.mjs");
+		const { runSetup } = await import("../setup-DwFaZOPf.mjs");
 		await runSetup();
 	});
 	await program.parseAsync(process.argv);
@@ -526,5 +526,4 @@ main().then((keepAlive) => {
 	console.error(JSON.stringify({ error: message }));
 	process.exit(1);
 });
-//#endregion
 export {};