The Agent Skills Directory

[PROMPT_INJECTION] (HIGH): The skill is highly vulnerable to indirect prompt injection. It is designed to read and analyze external code files (Ingestion Point: implicit in purpose) and has access to powerful tools like Bash, Write, and Edit (Capability Inventory). Malicious instructions embedded in comments or strings within the code being processed could be interpreted as commands by the agent, leading to unauthorized actions.
[COMMAND_EXECUTION] (MEDIUM): The inclusion of the Bash tool in the allowed-tools list provides a direct mechanism for command execution. While necessary for some automation tasks, it represents a significant security boundary risk when combined with untrusted inputs.
[NO_CODE] (INFO): This skill contains no executable scripts or logic in its definition. It is a configuration file that relies on the base agent's reasoning and the requested tools to perform its stated purpose.

jest-test-generator