baoyu-post-to-x

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • COMMAND_EXECUTION (MEDIUM): The skill executes system-level commands to perform browser and OS automation tasks.
  • Evidence: scripts/paste-from-clipboard.ts executes osascript (macOS), powershell.exe (Windows), and xdotool/ydotool (Linux).
  • Evidence: scripts/x-browser.ts and others spawn browser processes with --remote-debugging-port and other automation flags.
  • Evidence: scripts/x-utils.ts uses spawnSync to run local scripts via npx -y bun.
  • DATA_EXPOSURE (LOW): The skill creates and uses a persistent Chrome user data directory at ~/.local/share/x-browser-profile to maintain login sessions. This directory contains sensitive session cookies and local storage data.
  • DYNAMIC_EXECUTION (MEDIUM): Uses the Chrome DevTools Protocol (CDP) Runtime.evaluate command to inject and execute JavaScript within the context of the X website for DOM manipulation.
  • INDIRECT_PROMPT_INJECTION (LOW): The skill processes untrusted user data (text and Markdown articles) and injects it into a web browser session.
  • Ingestion points: User-provided text content and Markdown files in x-browser.ts and x-article.ts (referenced).
  • Boundary markers: Absent; no delimiters are used to separate user content from the automation scripts.
  • Capability inventory: Process spawning (shell commands), browser automation (CDP), and system clipboard manipulation.
  • Sanitization: Minimal; uses basic stringification before injecting content into the browser's JavaScript environment.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:25 PM