Skip to main content
skill CX Codex

faebryk — Claude Code skill: what it does, how to install

This skill gives an agent context for faebryk’s TypeGraph, edge traversal, references, and FabLL mappings. Use it when changing the graph core or its queries.

MIT 1 installs ★ 4,090 updated 2026-10-14

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add atopile/atopile --skill faebryk -a claude-code
Open on GitHub

What faebryk does

faebryk guides the agent through TypeGraph and GraphView, including the Zig edge implementation exposed to Python through faebryk.core.faebrykpy. It covers type nodes versus instance nodes, reference paths built with EdgeTraversal, and the meaning of edges such as EdgeComposition, EdgeTrait, EdgePointer, and EdgeOperand.

The material points to src/faebryk/core/faebrykpy.py, the generated typegraph.pyi stub, and src/faebryk/core/zig/src/faebryk/*. It also connects the graph to FabLL, compiler stages, solver traversal, and the build/export pipeline. For trait queries, it recommends the typed EdgeTrait.traverse(trait_type=SomeTrait) form instead of a string-based trait name.

Zig changes require ato dev compile to rebuild the bindings; otherwise the Python layer will not receive the updated implementation. The listed checks cover TypeGraph, runtime behavior, and Zig-backed traversal with ato dev test --llm.

What it helps with

  • trace references through TypeGraph
  • distinguish composition, pointer, and trait edges
  • map FabLL traits to graph edge types
  • build type-aware graph queries
  • run checks for Zig-backed traversal

Example prompts for the agent

  • Explain how GraphView, TypeGraph, and instance nodes relate to each other.
  • Show a typed trait traversal with EdgeTrait.traverse(trait_type=SomeTrait).
  • Review changes under src/faebryk/core/zig/src/faebryk/* and choose the relevant tests.

How to install faebryk

1Skills CLI (Vercel)

npx skills add atopile/atopile --skill faebryk -a claude-code

2Manually: put the files in place

~/.claude/skills/faebryk/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/faebryk/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about faebryk

How do I install faebryk in Claude Code?

Run npx skills add atopile/atopile --skill faebryk -a claude-code, then check ~/.claude/skills/faebryk/SKILL.md.

What does faebryk do?

It helps an agent understand TypeGraph structure, follow references, select edge semantics, and connect FabLL types and traits with the graph.

What should I do after changing the Zig implementation?

Run ato dev compile after editing src/faebryk/core/zig/src/faebryk/* so the Python bindings are rebuilt.

Source text

SKILL.md · MIT
---
name: faebryk
description: "How Faebryk's TypeGraph works (GraphView + Zig edges), how to traverse/resolve references, and how FabLL types/traits map onto edge types. Use when working with TypeGraph traversal, edge types, or building type-aware queries."
---

# Faebryk Core Module

The Faebryk core here is the **TypeGraph** + edge types implemented in Zig and exposed to Python via `faebryk.core.faebrykpy`.

Source-of-truth for API + behavior:
- `src/faebryk/core/faebrykpy.py` (Python-facing wrapper + type-safe `EdgeTrait.traverse`)
- `src/faebryk/core/zig/gen/faebryk/typegraph.pyi` (public stubbed API surface)
- `src/faebryk/core/zig/src/faebryk/*` (Zig implementation)

## Quick Start

```python
import faebryk.core.faebrykpy as fbrk
import faebryk.core.graph as graph

g = graph.GraphView.create()
tg = fbrk.TypeGraph.create(g=g)
```

## Relevant Files

- `src/faebryk/core/faebrykpy.py` (re-exports + `EdgeTraversal` + type-safe `EdgeTrait.traverse`)
- `src/faebryk/core/zig/gen/faebryk/typegraph.pyi` (TypeGraph stub)
- Key edge types (imported by `faebrykpy.py`):
  - `EdgeComposition` (parent/child structure)
  - `EdgeTrait` / `Trait` (trait attachment)
  - `EdgePointer` (references)
  - `EdgeInterfaceConnection` (interface connections)
  - `EdgeOperand` (solver operand wiring)
  - `EdgeType` / `EdgeNext` (type graph plumbing)
- Linker:
  - `Linker` (used by compiler/linking stages)

## Dependants (Call Sites)

- FabLL: `src/faebryk/core/node.py` (binds Python classes into the TypeGraph; uses composition/trait edges)
- Compiler: `src/atopile/compiler/*` (creates and links TypeGraphs)
- Solver: `src/faebryk/core/solver/*` (operand edges and instance traversal)
- Build/export pipeline: `src/atopile/build_steps.py` (visits type/instance edges for PCB/layout features)

## How to Work With / Develop / Test

### Core Concepts
- **GraphView + TypeGraph**: a `TypeGraph` is created against a `GraphView`:
  ```python
  import faebryk.core.graph as graph
  import faebryk.core.faebrykpy as fbrk

  g = graph.GraphView.create()
  tg = fbrk.TypeGraph.create(g=g)
  ```
- **Type nodes vs instance nodes**:
  - TypeGraph stores type definitions (“what exists structurally on a type”)
  - GraphView also holds instances created from those types (“a concrete design graph”)
- **EdgeTraversal**: `TypeGraph.ensure_child_reference(..., path=[...])` uses `EdgeTraversal` items to walk references through the type graph.

### Development Workflow
1) Zig-side changes: edit `src/faebryk/core/zig/src/faebryk/*` (edges, typegraph internals).
2) Rebuild bindings: `ato dev compile` (imports `faebryk.core.zig`).
3) Python ergonomics: add wrappers/helpers in `src/faebryk/core/faebrykpy.py` (example: type-safe `EdgeTrait.traverse`).

### Testing
- TypeGraph-heavy tests live in compiler/runtime suites:
  - `ato dev test --llm test/compiler/test_typegraph.py -q`
  - `ato dev test --llm test/compiler/test_runtime.py -q`
- Zig-backed traversal tests:
  - `ato dev test --llm test/core/zig/test_interface_pathfinder.py -q`

## Best Practices
- Import edges/TypeGraph via `faebryk.core.faebrykpy` (so callers get Python helpers, not just raw generated types).
- Prefer type-safe trait traversal:
  - `EdgeTrait.traverse(trait_type=SomeTrait)` over stringly-typed `trait_type_name=...`.
- When building reference paths, be explicit about edge semantics (composition vs pointer vs trait) rather than relying on implicit string behavior.

© atopile · MIT · atopile/atopile

Similar skills

All skills for Programming languages →

From the same set2

Project sponsors

Companies that keep this analytics open