The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to activate a virtual environment (source ../../.venv/bin/activate) and run local Python scripts (python ../findata-toolkit-cn/scripts/views_runner.py). This creates a functional dependency on an external local directory structure.
[EXTERNAL_DOWNLOADS]: The skill performs dependency installation using pip install -r ../findata-toolkit-cn/requirements.txt, which fetches third-party packages from external registries.
[PROMPT_INJECTION]: The skill exhibits an indirect command injection surface where user-supplied inputs (e.g., stock symbols, date windows) are interpolated directly into command-line arguments using the --set key=value syntax. Maliciously crafted inputs containing shell metacharacters could potentially attempt to escape the command context. \n- Ingestion points: User-provided parameters for '标的/范围' (target) and '时间窗口' (time window) specified in SKILL.md. \n- Boundary markers: The skill does not define specific delimiters or use instructions to ignore embedded commands within user data. \n- Capability inventory: The agent has the capability to execute shell commands and Python scripts as defined in references/data-queries.md. \n- Sanitization: No explicit input validation or escaping mechanisms are described in the provided methodology or data query files to pre-process user strings before they are used as CLI arguments.

ipo-lockup-risk-monitor