Linting And Formatting

When to use this skill

Use Linting And Formatting when an AI agent needs a reusable SKILL.md workflow for this job: Code style and quality rules for NeMo AutoModel — ruff configuration, naming conventions, type hints, docstrings, copyright headers, and the code review checklist.

When not to use it

Skip Linting And Formatting when the task is outside the coding category, or when a more specific skill in this directory already covers the same workflow with clearer triggers.

How to install

1. Personal install: create ~/.claude/skills/linting-and-formatting/SKILL.md (and any bundled scripts) so Claude Code, Claude Desktop, and compatible agents can load it in every project.
2. Project install: commit the same folder at .claude/skills/linting-and-formatting/ so teammates get the skill with the repo.
3. Restart the agent session after copying files so it re-scans the skills directory, then ask for the task in words that match the skill description.

Full install guide for Claude, Cursor, and Codex

What this skill does

# Linting and Formatting

Single source of truth for code style in NeMo AutoModel. Read this before writing new code or reviewing PRs.

## Formatting and Linting

Run before every commit:

```bash # Auto-format all source files (line length, quotes, trailing commas, etc.) ruff format .

# Lint and auto-fix what can be fixed (unused imports, isort, etc.) ruff check --fix . ```

To check without modifying files:

```bash ruff format --check . # exits non-zero if any file would change ruff check . # exits non-zero on lint violations bandit -r app.py nemo_automodel examples scripts tools tutorials -t B614 ```

To lint a single file or directory:

```bash ruff format nemo_automodel/components/models/llama/model.py ruff check --fix nemo_automodel/components/models/llama/ ```

### What linting enforces (from `pyproject.toml`)

| Rule | ID | Description | |---|---|---| | Line length | — | 120 characters (formatter) | | Quote style | — | Double quotes | | Unused imports | F401 | Auto-removed by `--fix` (ignored in `__init__.py`) | | Unused variables | F841 | Auto-removed by `--fix` | | Undefined names | F821 | Error | | f-string without placeholders | F541 | Error | | Import sorting | I | isort-compatible ordering, auto-fixed | | Docstring convention | D101/D103 | Google style (currently ignored — selected then suppressed) | | No pickle | S301/S403 | Security: forbids `pickle.load` | | Restricted PyTorch load | B614 | Requires explicit `weights_only=True` for `torch.load` | | Ambiguous variable names | E741 | Error (e.g., `l`, `O`, `I`) |

Tests (`tests/`) are excluded from Ruff and Bandit checks. Docstring rules (`D`) are also relaxed in test files.

## Type Hints

Required on all public API functions and methods.

- Use `T | None` instead of `Optional[T]` - Use `X | Y` instead of `Union[X, Y]` - Use built-in generics (`list`, `dict`, `tuple`) instead of `typing` equivalents

## Docstrings

Google-style where docstrings are added:

```python def build_model(config: dict) -> torch.nn.Module: """Instantiate and shard the model from config.

Args: config: Mapping with _target_ and model hyperparameters.

Returns: Fully initialized model ready for distributed training. """ ```

## NVIDIA Copyright Header

Every Python file must start with the NVIDIA copyright block. Do not remove or modify it. Use the current year (2026).

```python # Copyright (c) 2026, NVIDIA CORPORATION. All rights reserved. # # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # # http://www.apache.org/licenses/LICENSE-2.0 # # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. ```

## Additional Style Conventions

- **Explicit over implicit.** Inline logic where possible; avoid hiding behavior behind unnecessary layers of indirection. - **No speculative abstractions.** Do not add features, parameters, or generalization beyond what is explicitly asked for. - **Optional dependencies** must be guarded with `safe_import()` from `nemo_automodel.shared.import_utils`. Never let an optional import crash module loading. - **Components must not import each other** — enforced by `import-linter` (see `pyproject.toml`).

## Code Review Checklist

1. **Copyright header** present on all new Python files 2. **Type hints** on all public functions and methods 3. **Docstrings** on public classes and functions (Google style) 4. **Double quotes** for strings 5. **No bare `print()`** — use `logging.getLogger(__name__)` 6. **No commented-out code** without explanation 7. **Optional imports** guarded with `safe_import()` 8. **No cross-component imports** between `components/` subdirectories

## Automated Review

The review-only maintainability heuristics and thresholds live in `.github/workflows/claude-review.yml`. Keep repository-wide coding rules here and automated-review prompt policy there so the detailed checklist has one source of truth.

Intended uses

• Use Linting And Formatting when this documented workflow matches the task.