add-placeholder

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run mvn clean package -pl core -am -q. This is a standard Maven command used to verify that the code modifications compile correctly within the project. It is restricted to the specific project module.
  • [PROMPT_INJECTION]: The skill is designed to modify source code based on user requests, which creates a surface for indirect prompt injection.
  • Ingestion points: User requests such as 'add placeholder' or 'new variable' received at runtime.
  • Boundary markers: The skill includes a 'Critical' section that strictly limits modifications to a single file (PlaceholderReplacer.java) and forbids altering existing safety guards (null-checks).
  • Capability inventory: The skill utilizes file system access to read and write Java source code and the capability to execute the Maven build tool.
  • Sanitization: The skill enforces the use of specific Java code templates and fallback literals to maintain code integrity, though it does not explicitly sanitize the user-provided placeholder names.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 07:27 PM