test-driven-generation

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to run test suites.
  • Evidence: Step 3 ('Run Tests') lists specific commands for Python (pytest, unittest) and Java (mvn, gradle, javac, java).
  • [PROMPT_INJECTION]: The skill contains an attack surface for indirect prompt injection (Category 8) as it processes and executes code from user-provided files.
  • Ingestion points: The agent reads and processes user-provided test files (e.g., <test_file>.py, <TestFile>.java) in Step 1 and Step 3 of SKILL.md.
  • Boundary markers: No explicit boundary markers or instructions to treat user-provided file content as untrusted are present in the skill.
  • Capability inventory: The skill utilizes subprocess execution of test runners (pytest, java, etc.) which can execute arbitrary code contained within the provided test files.
  • Sanitization: No sanitization or validation of the user-provided code is performed before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 10:21 PM