The Agent Skills Directory

[COMMAND_EXECUTION]: The skill workflow requires the agent to execute shell commands including ./mvnw compile, mvn compile, ./mvnw clean verify, and mvn clean verify (SKILL.md). These are standard development operations for Java projects to verify that code changes do not break compilation or tests.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external Java source code provided by the user (SKILL.md, Workflow step 2). This creates a surface where maliciously crafted comments or code structures could attempt to influence the agent's analysis, although the risk is minimized by the skill's structured workflow and reliance on internal reference files.
Ingestion points: Affects the analysis of Java code provided in the user request or file system (SKILL.md).
Boundary markers: The agent is directed to use specific, internal reference files for implementation guidance, which serves as a control against external influence.
Capability inventory: The agent has the ability to execute Maven commands and modify local Java files.
Sanitization: The skill relies on the underlying model's safety guardrails when processing external code content.

121-java-object-oriented-design