The Agent Skills Directory

[COMMAND_EXECUTION]: Path Traversal Vulnerability. The scripts/download_report.py script constructs the output file path by joining a directory with a filename built from stock_code and year without sanitization. An attacker could use directory traversal sequences (e.g., ../) to write PDF files to arbitrary locations on the file system.
[COMMAND_EXECUTION]: Shell Command Surface. The skill executes a Python script using arguments derived from external web search results and user input. This represents a surface for potential command injection if the agent or execution environment does not properly escape these inputs before they are executed in a shell.
[EXTERNAL_DOWNLOADS]: The skill downloads PDF documents from stockn.xueqiu.com and notice.10jqka.com.cn. The script includes domain validation and checks for PDF magic bytes (%PDF-), which mitigate the risk of downloading non-PDF payloads. It also installs the requests library via pip.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill ingests and acts upon search results from external financial platforms. Ingestion points: WebSearch results for stock report titles and URLs in SKILL.md. Boundary markers: None. Capability inventory: Shell command execution, network access (downloads), and local file writing. Sanitization: The Python script performs domain whitelisting and file header verification.

report-download