The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to command injection because it interpolates untrusted user data (class names, namespaces, and file paths) directly into a shell command executed via execute_tool. While the documentation advises wrapping values with spaces in single quotes, it fails to provide guidelines for sanitizing or escaping shell-active metacharacters such as semicolons (;), pipes (|), ampersands (&), or backticks (`). An attacker could provide a crafted input that breaks out of the intended command to execute arbitrary shell operations.
Ingestion points: User-supplied strings for --className, --classNamespace, --filePath, --methodName, and --methodArguments in SKILL.md.
Boundary markers: The instructions recommend using single quotes for parameters containing spaces.
Capability inventory: The execute_tool function is used to invoke a system-level CLI tool (findTests).
Sanitization: There is no evidence of input validation, sanitization, or escaping for shell metacharacters beyond simple space-quoting.
[COMMAND_EXECUTION]: The skill's primary function is the execution of a system command (findTests) through the execute_tool capability. While this is the intended design for a JetBrains development tool, the combination of CLI execution and the aforementioned lack of input sanitization constitutes a security risk.

finding-tests