xiaohongshu-upload

Pass

Audited by Gen Agent Trust Hub on Oct 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The primary function of this skill is to execute the sau (social-auto-upload) command-line interface. Documentation and example scripts in scripts/examples/ provide instructions for performing account logins, cookie checks, and media uploads through shell execution.
  • [INDIRECT_PROMPT_INJECTION]: The skill interpolates user-controlled data such as titles, descriptions, and notes into CLI commands, creating a potential injection surface.
  • Ingestion points: Metadata fields such as --title, --desc, and --note defined in references/cli-contract.md and used in SKILL.md.
  • Boundary markers: CLI commands use double quotes to wrap user-provided strings.
  • Capability inventory: The skill utilizes subprocess.run (Python), PowerShell, and Bash to execute these commands across various scripts.
  • Sanitization: The Python template in scripts/examples/xiaohongshu_cli_template.py follows best practices by using list-based subprocess.run (avoiding shell=True) and uses shlex.quote for logging outputs.
  • [EXTERNAL_DOWNLOADS]: The references/runtime-requirements.md file references npmmirror.com for downloading browser binaries. This is a well-known mirror service maintained by Alibaba Cloud and is used here for standard dependency installation.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 24, 2026, 04:53 PM