fix: security audit round 2 (v0.13.4.0) (#640)

* fix: chrome-cdp localhost-only binding

Restrict Chrome CDP to localhost by adding --remote-debugging-address=127.0.0.1
and --remote-allow-origins to prevent network-accessible debugging sessions.

Clears 1 Socket anomaly (Chrome CDP session exposure).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: extension sender validation + message type allowlist

Add sender.id check and ALLOWED_TYPES allowlist to the Chrome extension's
message handler. Defense-in-depth against message spoofing from external
extensions or future externally_connectable changes.

Clears 2 Socket anomalies (extension permissions).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: checksum-verified bun install

Replace unverified curl|bash bun installation with checksum-verified
download-then-execute pattern. The install script is downloaded, sha256
verified against a known hash, then executed. Preserves the Bun-native
install path without adding a Node/npm dependency.

Clears Snyk W012 + 3 Socket anomalies.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: content trust boundary markers in browse output

Wrap page-content commands (text, html, links, forms, accessibility,
console, dialog, snapshot) with --- BEGIN/END UNTRUSTED EXTERNAL CONTENT ---
markers. Covers direct commands (server.ts), chain sub-commands, and
snapshot output (meta-commands.ts).

Adds PAGE_CONTENT_COMMANDS set and wrapUntrustedContent() helper in
commands.ts (single source of truth, DRY). Expands the SKILL.md trust
warning with explicit processing rules for agents.

Clears Snyk W011 (third-party content exposure).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* fix: harden trust boundary markers against escape attacks

- Sanitize URLs in markers (remove newlines, cap at 200 chars) to prevent
  marker injection via history.pushState
- Escape marker strings in content (zero-width space) so malicious pages
  can't forge the END marker to break out of the untrusted block
- Wrap resume command snapshot with trust boundary markers
- Wrap diff command output with trust boundary markers
- Wrap watch stop last snapshot with trust boundary markers

Found by cross-model adversarial review (Claude + Codex).

* chore: bump version and changelog (v0.13.4.0)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* chore: gitignore .factory/ and remove from tracking

Factory Droid support was removed in this branch. The .factory/ directory
was re-added by merging main (which had v0.13.5.0 Factory support).
Gitignore it so it stays out.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

browse/src/commands.ts

@@ -40,6 +40,21 @@ export const META_COMMANDS = new Set([
 
 export const ALL_COMMANDS = new Set([...READ_COMMANDS, ...WRITE_COMMANDS, ...META_COMMANDS]);
 
+/** Commands that return untrusted third-party page content */
+export const PAGE_CONTENT_COMMANDS = new Set([
+  'text', 'html', 'links', 'forms', 'accessibility',
+  'console', 'dialog',
+]);
+
+/** Wrap output from untrusted-content commands with trust boundary markers */
+export function wrapUntrustedContent(result: string, url: string): string {
+  // Sanitize URL: remove newlines to prevent marker injection via history.pushState
+  const safeUrl = url.replace(/[\n\r]/g, '').slice(0, 200);
+  // Escape marker strings in content to prevent boundary escape attacks
+  const safeResult = result.replace(/--- (BEGIN|END) UNTRUSTED EXTERNAL CONTENT/g, '--- $1 UNTRUSTED EXTERNAL C\u200BONTENT');
+  return `--- BEGIN UNTRUSTED EXTERNAL CONTENT (source: ${safeUrl}) ---\n${safeResult}\n--- END UNTRUSTED EXTERNAL CONTENT ---`;
+}
+
 export const COMMAND_DESCRIPTIONS: Record<string, { category: string; description: string; usage?: string }> = {
   // Navigation
   'goto':    { category: 'Navigation', description: 'Navigate to URL', usage: 'goto <url>' },