What is pypsa-skills-kit?

PyPSA Skills Kit is a collection of nine knowledge files that teach AI coding assistants how to work with PyPSA, an open-source Python library for modeling and optimizing energy systems (power grids, batteries, hydrogen networks, and similar infrastructure). When engineers use tools like Claude Code or Windsurf to build energy models, the AI knows the programming API but not the domain expertise: which modeling choice avoids a subtle error, why a particular revenue estimate will be too optimistic before a lender ever sees it, or what a suspicious shadow price actually signals. These skill files encode that judgment in a form the AI can load and apply. The kit contains nine skills, each answering a different class of question. Some are diagnostic: one helps debug a model that refuses to solve, another checks whether a solved result is physically plausible. Some are design-focused: one covers how to represent heat, hydrogen, transport, and industrial loads in a single model, another explains market design choices like nodal versus zonal pricing and how congestion revenue works. Two address economics: one flags foresight bias that makes storage revenue look artificially high, another handles the practical question of where to get realistic input data for fuel prices, technology costs, and weather. The skill files are plain Markdown. Installation means copying them into the folder your AI tool watches for skills or rules files. Once installed, the right skill loads automatically when you describe a problem, or you can call one explicitly by name. The full kit is about 23,000 tokens of content, but a typical session loads only around 2,600 tokens because most skills stay on disk until needed. The author ran every code example in the files against actual solved networks before publishing, which caught at least two errors (a sign inversion and an arbitrage trap) that would otherwise have quietly produced wrong answers. A smoke test script compiles all scripts, solves a synthetic network, and runs the validator, so the test suite exercises real computation rather than just checking that the files parse. The kit is compatible with PyPSA 1.0.7, the HiGHS solver, and the linopy constraint language. It also supports workflow frameworks built on top of PyPSA, including PyPSA-Eur and PyPSA-Earth. The license is MIT.