The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions promote the interpolation of user-supplied input into shell command templates. Specifically, the pattern python "<SKILL_DIR>/../../scripts/look_tongji.py" list --query "<KEYWORD>" relies on the agent to safely handle the <KEYWORD> placeholder. If a user provides input containing shell metacharacters (e.g., "; rm -rf /; #), it could lead to arbitrary command execution on the host system.
[SAFE]: The skill interacts with the domain look.tongji.edu.cn, which is the official platform for Tongji University. This network activity is consistent with the skill's stated purpose of discovering courses.
[SAFE]: The skill presents a surface for indirect prompt injection.
Ingestion points: Data is fetched from the external look.tongji.edu.cn server (course titles, teacher names).
Boundary markers: None are defined in the instructions to separate external data from agent instructions.
Capability inventory: The agent has the capability to execute local Python scripts (look_tongji.py).
Sanitization: No sanitization or validation of the fetched course data is described before it is processed by the agent. If an attacker controls course metadata on the server, they could attempt to influence agent behavior.

list