The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill takes user input to determine the endpoint for smoke tests, which is then passed as an argument to a scaffolding script.
Ingestion points: Natural language user input is used to generate the command line for init_restassured_module.py (as seen in SKILL.md examples).
Boundary markers: No specific delimiters or safety instructions are used to wrap user-provided endpoint paths when the agent constructs the command.
Capability inventory: The init_restassured_module.py script possesses the capability to write Java source files, build configurations (Maven/Gradle), and properties to the filesystem using Path.write_text.
Sanitization: The script scripts/init_restassured_module.py does not sanitize the smoke_path argument before injecting it into the SMOKE_TEST_TEMPLATE. This lack of escaping could allow for the generation of malformed Java code if the input contains double quotes or other Java delimiters.
[COMMAND_EXECUTION]: The skill requires the agent to execute a local Python script (init_restassured_module.py) to perform its primary function, which involves automated file creation and directory modification on the host system.
[DYNAMIC_EXECUTION]: The skill generates Java source code and build files (pom.xml, build.gradle.kts) from internal templates at runtime. The resulting code is intended for subsequent execution by the user, representing a template-based code generation workflow.

restassured-bootstrap