An AI agent skill for working with Entity, an open-source general-relativistic particle-in-cell (PIC) plasma simulation code.
Teaches AI coding agents the full Entity simulation workflow:
- Build — check environment, install dependencies, configure CMake, compile
- Configure — write TOML input files, edit problem generators (pgen.hpp)
- Run — launch simulations, interpret output, handle checkpoints
- Debug — diagnose compile/runtime errors, optimize performance
- Analyze — load and visualize simulation data with nt2py
Copy the entire entity-toolkit-skill/ folder into your agent's skills directory.
ln -s $(pwd)/entity-toolkit-skill .claude/skills/entity-toolkit-skillCopy the folder to wherever your agent loads skills from. The skill is self-contained — all knowledge file references use relative paths.
entity-toolkit-skill/
├── SKILL.md # Entry point
├── README.md
└── knowledge/
├── entity/
│ ├── 01-build-env.md # CMake options, GPU flags, compile
│ ├── 01b-install-methods.md # Dependencies: pip, Docker, Spack, manual
│ ├── 02-toml-config.md # Complete TOML parameter reference
│ ├── 03-pgen-guide.md # Problem generator C++ API guide
│ └── 04-debug.md # Compile & runtime error diagnosis
└── nt2py/
├── 01-data-loading.md # Installation, output files, Data API
├── 02-fields.md # xarray field plotting & movies
└── 03-particles-stats.md # Phase-space, spectra, diagnostics
The skill assumes the agent has access to:
- The Entity source code (clone from
entity-toolkit/entity) - A build environment with CMake and a C++20 compiler
- Python 3.7+ with
nt2pyfor analysis
BSD-3-Clause