300-frameworks-spring-boot-create-project
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute multiple shell commands to manage environment configurations and project build tasks. This includes
sdk version,java -version,spring init, and./mvnw clean verify. While these are legitimate development tasks, they involve direct interaction with the host system's shell. - [EXTERNAL_DOWNLOADS]: The skill utilizes SDKMAN (
sdk install) to download and install Java and Spring Boot CLI binaries. Additionally, thespring initcommand downloads project templates, typically fromstart.spring.io. While these sources are standard in the Java ecosystem, they involve the retrieval and execution of external code. - [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection and command injection because user-supplied strings (such as group IDs, artifact IDs, and dependency lists) are interpolated directly into a shell command (
spring init ... --group-id=<group> --artifact-id=<artifact> ...). - Ingestion points: Project metadata fields (group, artifact, name, package, dependencies) provided by the user in
references/300-frameworks-spring-boot-create-project.md. - Boundary markers: None. User-supplied values are placed directly into the command string without delimiters or escaping.
- Capability inventory: Execution of
spring,sdk, and the generated Maven wrapper (./mvnw) via the shell. - Sanitization: The instructions do not specify any validation or sanitization requirements for the user-supplied project coordinates before they are used in command execution.
Audit Metadata