soia-pkm-clip-wechat-account

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXFILTRATION]: The skill communicates with api.weixin.qq.com and mp.weixin.qq.com. These are official domains for the WeChat service and are necessary for the skill's primary function of fetching article content. No sensitive user data is sent to unauthorized or unknown third-party domains.
  • [CREDENTIALS_UNSAFE]: The skill requires sensitive credentials (API keys, cookies) to operate. It implements secure management practices by loading these secrets from a local configuration file in the user's home directory (~/.config/soia-skills/) rather than hardcoding them or requiring them in the skill's source code.
  • [COMMAND_EXECUTION]: The skill is implemented in pure Python using standard library modules for networking and file operations. It does not utilize subprocess, os.system, or similar functions to execute arbitrary shell commands. It includes proper escaping using shlex.quote when assisting users with environment exports.
  • [EXTERNAL_DOWNLOADS]: All external requests are directed toward official WeChat API endpoints and public article pages. The skill does not download executable code, scripts, or binary assets from untrusted sources.
  • [PROMPT_INJECTION]: No instructional patterns designed to bypass AI safety filters or override agent constraints were detected. The skill's Markdown conversion logic for ingested articles includes basic sanitization by stripping <script> and <style> tags, reducing the surface for indirect injection attacks.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 08:06 AM
Security Audit — agent-trust-hub — soia-pkm-clip-wechat-account