The Agent Skills Directory

[COMMAND_EXECUTION]: The script vensim_system_dynamics/tools/vensim_autolayout.py utilizes subprocess.run() to invoke Graphviz binaries for calculating model layouts. This was analyzed and determined to be safe as the binary name is validated against a strict whitelist of engines (dot, neato, fdp, sfdp) and arguments are passed as a list, preventing shell injection risks.
[SAFE]: The simulation engine in vensim_system_dynamics/tools/vensim_engine.py evaluates mathematical expressions from model files using the Python ast module. The implementation uses a custom walker (_safe_eval_node) that strictly limits execution to authorized mathematical functions and operations, successfully mitigating risks associated with dynamic code execution.
[DATA_EXFILTRATION]: No network access or data exfiltration capabilities were found within the provided scripts. The skill operates entirely on local .mdl and configuration files.
[CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or other sensitive credentials were found in the skill metadata or source code.

vensim-skill