The Agent Skills Directory

[COMMAND_EXECUTION]: The pipeline orchestration scripts xuexitong_hw_pipeline.py and xuexitong_scan_pending.py utilize subprocess.run and subprocess.check_output to execute helper modules like the rendering script and the submission script. These calls are implemented using list-based arguments, which follows security best practices to prevent shell command injection.
[EXTERNAL_DOWNLOADS]: The update_check.py script performs a network request to the project's GitHub repository to fetch the latest version string. This is a common and benign feature used for update notifications and does not involve the direct execution of downloaded remote code.
[DATA_EXFILTRATION]: The skill accesses authentication cookies from a dedicated local file (~/.openclaw/credentials/xuexitong_cookie.txt) to authenticate the automated requests. These credentials are transmitted to legitimate platform endpoints (mooc1-api.chaoxing.com and notice.chaoxing.com) as required for the skill's primary function.
[INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it ingests untrusted content from the external Xuexitong platform.
Ingestion points: Assignment pages and task URLs fetched in xuexitong_submit.py.
Boundary markers: None identified in the prompt interpolation logic.
Capability inventory: Subprocess command execution, file system writes, and network POST operations.
Sanitization: Employs basic regular expression filtering to clean HTML tags from question stems before processing.

xuexitong-homework-submit