create-test-project

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8) by using user-supplied specifications to generate executable Java source files and build configurations.
  • Ingestion points: The <spec> input (defining the test requirements) and the <tracking-file> path entering the agent context from the caller in SKILL.md.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when interpolating the specification into the source generation process.
  • Capability inventory: The skill possesses the ability to execute shell commands (opentaint, javap) and write arbitrary files to the local filesystem (Java samples in <test-project> and YAML updates in <tracking-file>).
  • Sanitization: The instructions do not specify any validation or sanitization for the input text or file paths provided by the user.
  • [COMMAND_EXECUTION]: The skill invokes the opentaint CLI for project initialization and compilation (e.g., opentaint test rule init, opentaint compile). These commands are standard operations for the tool associated with the skill author.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 10:39 AM
Security Audit — agent-trust-hub — create-test-project