mock-interview

Fail

Audited by Gen Agent Trust Hub on Nov 10, 2026

Risk Level: HIGHCOMMAND_EXECUTIONDYNAMIC_EXECUTIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill implements Text-to-Speech (TTS) functionality in VOICE.md by interpolating interviewer dialogue directly into shell commands such as say, spd-say, and PowerShell's SpeechSynthesizer. Because the dialogue often includes fragments of the candidate's input (e.g., repeating variable names, approach titles, or answering specific user questions), a malicious user could inject shell metacharacters to execute unauthorized commands on the host system.
  • [DYNAMIC_EXECUTION]: In SKILL.md, the skill is instructed to build a grader that executes a reference solution, a brute-force solution, and the candidate's provided code within a temporary directory. Executing arbitrary code provided by an untrusted user (the candidate) is a high-risk operation. While the skill mentions using a "runner" with timeouts, it does not specify a secure sandbox environment, leaving the host vulnerable to exploitation by malicious code submitted as an "interview solution."
  • [COMMAND_EXECUTION]: The skill executes a Python script (questions.py) using parameters like <Company> and <level> which are derived from user requests or profiles. If a user provides a crafted company name (e.g., including shell pipes or semicolons), it could result in command injection during the problem selection phase.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input in the form of code, interview responses, and clarifying questions. This input is then used to determine the agent's next actions and is interpolated into the aforementioned shell commands. The lack of explicit sanitization or strict boundary markers for this untrusted data creates a vulnerability surface for indirect prompt injection or command injection attacks.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Nov 10, 2026, 06:30 AM