The Agent Skills Directory

[SAFE]: The skill implements robust input validation across all scripts. Configuration parameters are checked against a strict allowlist of characters, and numeric thresholds are validated as finite, positive floats to prevent unexpected behavior from malformed input.
[SAFE]: File access is managed securely. Log parsing scripts (runtime_monitor.py, failure_diagnoser.py) enforce a 500 MB file size limit before reading content, mitigating potential denial-of-service or memory exhaustion risks.
[SAFE]: The codebase avoids dangerous execution patterns. No instances of eval(), exec(), or unsafe subprocess calls were found. The use of ctypes in preflight_checker.py is restricted to calling a static, well-known Windows API (GetDiskFreeSpaceExW) for legitimate system monitoring.
[SAFE]: Log analysis logic relies on pre-compiled regular expressions with specific boundary anchors. This prevents false positives from domain-specific terminology and ensures that diagnostic data is interpreted correctly without risk of prompt or instruction injection through log content.

simulation-validator