defect: multiple execute_command_line calls pose security risks

[krystophny]

Found during PLAY workflow audit - security vulnerabilities in system command execution:

High-Risk System Command Calls

Found in fortplot_forensic_comparison.f90:

• Line 153: execute_command_line("git branch --show-current > current_branch.tmp")
• Line 174, 181, 197, 204: Multiple git and build commands without validation
• Direct shell command execution without input sanitization

Found in fortplot_matplotlib_io.f90:

• Line 257: execute_command_line("sleep 0.1", wait=.true.)
• Line 261: execute_command_line("sleep 1", wait=.true.)

Found in fortplot_security.f90:

• Line 220, 662: Commands executed with minimal validation

Security Vulnerabilities

1. Command Injection: No input sanitization before execution
2. Path Traversal: Insufficient path validation
3. Shell Escape: Direct shell access without sandboxing
4. Privilege Escalation: System commands run with process privileges

Impact Assessment

• HIGH: Forensic comparison module allows arbitrary command execution
• MEDIUM: Sleep commands hardcoded but could be modified
• MEDIUM: File operations without proper validation

Affected Functionality

• Forensic rendering comparison tools
• Image display functionality
• Test security operations
• System integration features

Recommended Fixes

1. Replace execute_command_line with secure alternatives
2. Implement command whitelist/validation
3. Use Fortran file I/O instead of shell commands
4. Sandbox system operations
5. Remove forensic comparison module or secure it properly

Security Module Contradiction

The codebase has fortplot_security.f90 and fortplot_system_secure.f90 modules that provide secure alternatives, but unsafe execute_command_line calls still exist elsewhere.