Skip to main content
The server exposes a managed Python environment for quick analysis and scripting.

​
Ad-hoc code

Use execute_python_code for one-off probes and analysis.
  • Prefer this over python -c in run_command.
  • Only persist files in /workspace when you intentionally want a reusable artifact.
{"tool":"execute_python_code","arguments":{"code":"print('hello from inline python')","cwd":"/workspace","timeout":120.0}}

​
Existing scripts

Use execute_python_script only when a .py file already exists. 
{"tool":"execute_python_script","arguments":{"script_path":"/workspace/solve.py","args":"--remote","cwd":"/workspace","timeout":300.0}}

​
Package management

  • install_python_packages installs extra packages with uv.
  • list_python_packages lists what is already available.
{"tool":"install_python_packages","arguments":{"packages":"z3-solver capstone","upgrade":false}}
{"tool":"list_python_packages","arguments":{}}