Line Execution Checker
Warn
Audited by Gen Agent Trust Hub on Jun 7, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The C++ source code (
line_checker.cpp) contains a command injection vulnerability. In thefind_gcov_filefunction, thesource_fileargument is concatenated directly into a shell command:std::string cmd = "gcov " + source_file + " >/dev/null 2>&1"; system(cmd.c_str());. - Evidence: If a user provides a filename like
file.c; rm -rf /, the resulting command executed by the system would begcov file.c; rm -rf / >/dev/null 2>&1. - [PROMPT_INJECTION]: The skill instructions in
SKILL.mddirect the agent to run the compiled tool using user-provided input without sanitization. While the agent's natural language processing acts as a filter, the lack of input validation in the underlying tool creates an exploitable surface for indirect or direct prompt injection that leads to command execution.
Audit Metadata