The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute several shell commands, including curl, unzip, rm, and ./gradlew.
[EXTERNAL_DOWNLOADS]: The skill downloads a project template from https://start.spring.io/starter.zip. This is a well-known service for Spring Boot developers.
[INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to command injection through the interpolation of the ${input:projectName} variable into shell commands.
Ingestion points: The project name is accepted as user input and used in the curl and unzip commands in SKILL.md.
Boundary markers: No markers or warnings are used to sanitize or delimit the user input within the shell strings.
Capability inventory: The skill uses curl, unzip, and ./gradlew which provide a path to code execution if the input is manipulated.
Sanitization: No validation, escaping, or filtering is performed on the user-supplied projectName before it is passed to the shell.
[CREDENTIALS_UNSAFE]: The skill includes hardcoded placeholder passwords (rootroot) for Redis, PostgreSQL, and MongoDB configurations in application.properties and docker-compose.yaml.

create-spring-boot-kotlin-project