swift-quiz

Pass

Audited by Gen Agent Trust Hub on Feb 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface. Ingestion points: User-provided Swift code in 'Type D' quiz questions. Boundary markers: Absent; the instructions do not specify delimiters when passing user code to the Bash subagent. Capability inventory: Execution of code via the Bash subagent and the swift compiler. Sanitization: Absent; the skill directs the agent to evaluate raw user input.
  • [COMMAND_EXECUTION] (LOW): The skill utilizes a Bash subagent to run the swift command on user-provided code strings. While executing untrusted input is a security risk, the severity is categorized as LOW because this behavior is required for the skill's primary intended use-case of testing coding proficiency.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 20, 2026, 07:42 AM