mathLabmathLab/PINA/AGENTS.md

PINA

AGENTS.md8 starsChanged 4 months agoMITView on GitHub

A PyTorch library for solving differential equations with neural networks (PINNs). Skills under .opencode/skills/<name>/SKILL.md give step-by-step guidance. Load the relevant skill when the user's request matches:

Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository19 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- condition-setupSkill
- create-problemSkill
- create-skillSkill
- define-domainsSkill
- define-equationsSkill
- pina-workflowSkill
- select-modelSkill
- select-solverSkill
- select-trainerSkill
- condition-setupSkill
- create-problemSkill
- create-skillSkill
- define-domainsSkill
- define-equationsSkill
- pina-workflowSkill
- select-modelSkill
- select-solverSkill
- select-trainerSkill

The file

We split this AGENTS.md into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Very specific:
Never modify source code (pina/, tests/, docs/).
line 78

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.

Files

AGENTS.md

PINA — Physics-Informed Neural Architectures

A PyTorch library for solving differential equations with neural networks (PINNs).

Quick Reference

Workflow: Problem → Model → Solver → Trainer

problem = MyProblem()
problem.discretise_domain(n=256, mode="grid")
model = FeedForward(input_dimensions=2, output_dimensions=1)
solver = PINN(problem=problem, model=model)
trainer = Trainer(solver=solver, max_epochs=1000)
trainer.train()

Problem types

SpatialProblem, TimeDependentProblem, ParametricProblem, InverseProblem. Define output_variables, spatial_domain/temporal_domain, and conditions (dict of Condition).

Condition types

Condition(domain=..., equation=...) — physics residual on sampled domain
Condition(input=..., equation=...) — physics residual at fixed points
Condition(input=..., target=...) — supervised data
Condition(input=..., n_windows=..., unroll_length=...) — time series

Domains

CartesianDomain, EllipsoidDomain, SimplexDomain. Set ops: Union, Intersection, Difference, Exclusion. Discretise with problem.discretise_domain(n, mode) where mode is "grid", "random", "lh", "chebyshev".

Solvers

PINN, CausalPINN, SelfAdaptivePINN, CompetitivePINN, GradientPINN, RBAPINN, SupervisedSolver, AutoregressiveSolver. Ensembles via *EnsembleSolver variants.

Models

FeedForward, ResidualFeedForward, PirateNet, DeepONet, MIONet, FNO, KolmogorovArnoldNetwork, GraphNeuralOperator, SINDy, and more.

Equation zoo

PoissonEquation, HelmholtzEquation, BurgersEquation, AdvectionEquation, AllenCahnEquation, AcousticWaveEquation, DiffusionReactionEquation. Boundary: FixedValue, FixedGradient, FixedFlux, FixedLaplacian.

Key utilities

LabelTensor — torch.Tensor with named columns; index via .extract(["x", "y"])
Differential operators: grad, div, laplacian, advection — not cached, compute once
Trainer wraps lightning.pytorch.Trainer; handles DataModule, batching, device placement

Skills

Skills under .opencode/skills/<name>/SKILL.md give step-by-step guidance. Load the relevant skill when the user's request matches:

Skill	When to load
`pina-workflow`	User wants to solve a PDE/ODE or model a physical system — orchestrates the full session
`create-problem`	Defining a problem, setting up variables and outputs
`define-domains`	Creating and discretising domains
`define-equations`	Writing PDEs/ODEs or using the equation zoo
`condition-setup`	Binding equations to domains, data-driven conditions
`select-model`	Choosing or configuring a neural architecture
`select-solver`	Selecting the right solver for the problem
`select-trainer`	Configuring training parameters and callbacks
`create-skill`	User wants to create or edit a skill in this repo
`skill-sync-checker`	Auditing skill references against the codebase

Commands

pytest tests/ -x       # run tests
ruff check .           # lint

Rules

Never modify source code (pina/, tests/, docs/). Skills are read-only guidance.
Stay conversational — walk through the workflow naturally.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free