The Agent Skills Directory

[COMMAND_EXECUTION]: The instructions include running standard build tool commands such as mvn test and gradle test to verify test execution and coverage. This is a standard and expected operation for a developer-oriented unit testing skill.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the analysis of project source code to generate unit tests, which represents an attack surface where untrusted input from source files could influence agent behavior.
Ingestion points: The skill ingests local Java source code (specifically @Service classes) using Read, Glob, and Grep tools as part of its test generation workflow.
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" markers to prevent the agent from following directions potentially embedded in source code comments.
Capability inventory: The skill utilizes Write tools to generate new test files and Bash tools to execute them via the project's build system.
Sanitization: There are no explicit sanitization or validation steps described for the input source code before it is processed by the agent to generate test logic.

unit-test-service-layer