baoyu-post-to-wechat

Pass

Audited by Gen Agent Trust Hub on Nov 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes system commands across multiple platforms to automate browser interactions and environment setup. This includes using osascript on macOS for accessibility and clipboard management, powershell.exe on Windows for SendKeys and clipboard access, and ssh for dynamic port forwarding.
  • [REMOTE_CODE_EXECUTION]: Supports a 'remote-api' publishing method that executes ssh to establish a SOCKS5 tunnel, allowing local processes to route traffic through a remote server. Additionally, it generates and executes temporary Swift scripts on macOS to handle rich text and image data on the system clipboard.
  • [DYNAMIC_EXECUTION]: The script scripts/check-permissions.ts and scripts/copy-to-clipboard.ts dynamically generate and execute Swift and AppleScript code to interface with system-level APIs for clipboard operations and window management.
  • [INDIRECT_PROMPT_INJECTION]: As the skill is designed to process user-provided Markdown and HTML files for conversion and publishing, it has an inherent surface for indirect prompt injection. The skill converts external links to bottom citations and processes image placeholders, which are standard operations for its intended use case.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 2, 2026, 09:25 AM