The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands using user-provided strings for product and store names. Specifically, the git add and git commit commands interpolate variables that are not sanitized, which could allow a malicious user to execute arbitrary commands if the agent does not validate the input.
[INDIRECT_PROMPT_INJECTION]: The skill creates a vulnerability surface by ingesting untrusted user input and using it in shell operations. * Ingestion points: Product name and store name phrasing from the user (Steps 1, 2, and 3). * Boundary markers: None are provided to isolate the user input within the shell command. * Capability inventory: Access to the shell to run git commands (Step 4). * Sanitization: No sanitization or escaping of shell metacharacters is required by the instructions.

shopping-list-add