110-java-maven-best-practices

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill requires the agent to run ./mvnw validate, mvn validate, and ./mvnw clean verify. Executing the Maven wrapper (mvnw) involves running a script provided by the project being analyzed, which is standard behavior but involves local code execution.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its ingestion of untrusted project configuration data.
  • Ingestion points: The skill reads the root pom.xml and all identified child module pom.xml files.
  • Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious content within the XML files.
  • Capability inventory: The agent is authorized to execute terminal commands (mvn, ./mvnw) and perform file system modifications based on the project state.
  • Sanitization: No specific sanitization or filtering of the content read from the pom.xml files is implemented.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 10:21 PM