The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands including ./mvnw compile, mvn compile, and mvn clean verify. Since these commands rely on project-local files like mvnw and pom.xml, they could lead to arbitrary code execution if the project being audited is malicious or compromised.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and refactor Java source code from the user's project, which is untrusted external data. Malicious instructions hidden in code comments or strings could potentially influence the agent's behavior during the refactoring process.
Ingestion points: Java source files within the project workspace analyzed by the agent.
Boundary markers: None specified; the agent is instructed to read the project code directly.
Capability inventory: Shell command execution (mvn), local file modification (refactoring), and network access (implied via Maven dependency resolution).
Sanitization: No specific sanitization or validation of the ingested code is described.

128-java-generics