The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a REST API structure that accepts external data via controller endpoints, creating a potential surface for indirect prompt injection if the agent processes the resulting data without sufficient isolation.
Ingestion points: External data enters through PersonController.java via @RequestBody in create and update methods, and @RequestParam in the list method.
Boundary markers: The skill utilizes standard Spring Bean Validation (@Valid) and JSR-303 annotations (e.g., @NotBlank, @Email) to enforce data constraints.
Capability inventory: The skill performs database writes through PersonRepository and logging via SLF4J in PersonService.java.
Sanitization: The skill follows best practices by using Spring Data JPA (which uses parameterized queries to prevent SQL injection) and provides structured logging placeholders instead of string concatenation.
[SAFE]: The skill demonstrates high security awareness by providing a .gitignore that excludes .env and application-local.yml files, and the Dockerfile configuration correctly implements a non-root user (spring) to run the application, adhering to the principle of least privilege.

java-spring-boot-app