publishing-to-cnblogs

Pass

Audited by Gen Agent Trust Hub on Aug 2, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill automates article publishing to cnblogs (博客园). Analysis of all scripts and documentation confirms that all operations are directed at official cnblogs domains (cnblogs.com, i.cnblogs.com, upload.cnblogs.com).
  • [COMMAND_EXECUTION]: The skill requires the user to manually launch a Chrome instance with remote debugging enabled (--remote-debugging-port=9222) to manage authentication state. This is a standard and safe method for browser automation that avoids the need for hardcoded credentials or untrusted login wrappers.
  • [DATA_EXFILTRATION]: The get_cookies.py script exports cookies from the local debug Chrome instance to a local cookies.json file. These cookies are subsequently used by other scripts (publish.py, rehost_images.py) to authenticate requests to the cnblogs API. No evidence of data being sent to unauthorized third-party servers was found; network traffic is limited to the user's defined source URLs (during scraping) and cnblogs (during publishing).
  • [EXTERNAL_DOWNLOADS]: The rehost_images.py script downloads images from external URLs found in the source article to the local environment and then uploads them to cnblogs. This is the intended primary purpose of the skill to bypass image anti-leeching protections.
  • [REMOTE_CODE_EXECUTION]: No remote code execution patterns were detected. The skill uses well-known Python libraries (selenium, requests, markdownify, pyyaml) for its operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 2, 2026, 03:23 AM