How real projects brief Claude Code: the commands, conventions and warnings they put in CLAUDE.md.
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.
NVIDIACLAUDE.mdNVIDIA/kumo-engine/packages/kumo-ui-catalog/CLAUDE.md
@kumo-ai/kumo-ui-catalog — Kumo's shared component library and design system. Built on Radix UI + Tailwind CSS + CVA. Consumed in-tree by the Kumo UI. All UI primitives follow this pattern:
Hugging FaceCLAUDE.mdhuggingface/physics-intern-v0/CLAUDE.md
This is a multi-agent scaffolding system for autonomous scientific research in mathematics and theoretical physics. See README.md for overview, motivation, and design principles. GitHub Actions (.github/workflows/testing.yml) runs three commands on…
MicrosoftCLAUDE.mdmicrosoft/winml-cli/CLAUDE.md
Never hardcode model architecture names, node/operator names, input/output tensor names, layer naming patterns, or any model-specific logic. All solutions must be universal and architecture-agnostic. All testing uses pytest with code-generated results. Never create standalone test scripts, use LLM-generated expectations, or generate test data manually. Run the necessary pytest scope after every implementation or test revision (for example, affected test files/modules/cases). Run full suite uv run pytest tests/ only when explicitly requested, before merge/release, or when impact is broad/unclear. Never assume…
MicrosoftCLAUDE.mdmicrosoft/winml-cli/tests/CLAUDE.md
Inherits all rules from /CLAUDE.md. Additional test-specific rules below. Reference: /docs/pytest-best-practices.md - Place unit tests under tests/unit/<module>/ mirroring src/winml/modelkit/<module>/ - Place integration tests under tests/integration/, e2e under tests/e2e/ - Place cross-cutting CLI-surface tests (startup, import budget, arg parsing, command discovery, version/help output) under tests/cli/ — these don't mirror any single src/ module, so they don't fit under tests/unit/<module>/ - Put shared fixtures in the narrowest conftest.py that covers all consumers - Place tests that mirror a src/winml/modelkit/<module>/ directly under tests/…
MicrosoftCLAUDE.mdmicrosoft/monosize/CLAUDE.md
monosize is a CLI tool that measures bundle sizes for packages and monorepos and compares them against a stored baseline (typically main). This repo is itself a Yarn 4 + Nx monorepo of the CLI plus first-party adapter packages. The packages/ directory has three roles: Adapters are wired up by users in their own monosize.config.mjs; this repo does not consume itself end-to-end. When changing the BundlerAdapter / StorageAdapter shape, every adapter package must be updated in lockstep. Per-package and full-graph…
NVIDIACLAUDE.mdNVIDIA/nvrc/CLAUDE.md
Read ARCHITECTURE.md first to understand the security model, threat landscape, and design rationale before making changes. NVRC is a minimal init process (PID 1) for ephemeral confidential VMs with NVIDIA GPUs. It configures GPU resources, starts required daemons (nvidia-persistenced, nv-hostengine, etc.), and hands off to kata-agent. The goal is to eventually run NVRC as no_std. We are slowly transitioning by building hardened_std as our security-hardened std replacement. This enables: **no_std Transition Roadmap:** Runtime-writable config pattern: The rootfs image is read-only.…
MicrosoftCLAUDE.mdmicrosoft/DataFactory.MCP/CLAUDE.md
.NET MCP server for Microsoft Fabric Data Factory operations.
NVIDIACLAUDE.mdNVIDIA/DALI_extra/CLAUDE.md
DALI_extra is a data-only repository containing sample datasets used by NVIDIA DALI's test suite. It has no build system or runnable code. Large binary files (.mp4, .avi, .mkv, .tar, LMDB files, TFRecord files, etc.) are stored via Git LFS. Git LFS must be installed before cloning or the large files will be stubs. Set the DALI_EXTRA_PATH environment variable to point to this repository's root before running DALI tests: Before adding new files: - Verify no existing dataset can be used…
VercelCLAUDE.mdvercel/microfrontends/CLAUDE.md
This is a monorepo for @vercel/microfrontends, a library that enables splitting large applications into independently developed and deployed microfrontends on Vercel. The package provides generic platform support for microfrontends in multiple frameworks, such as Next.js, SvelteKit, React Router, Vite, and React. When running commands for a specific application, do not cd into the directory. Run commands from the root. The above commands can be run for a particular package by using a filter option: pnpm checks -F <package name>. Use…
JetBrainsCLAUDE.mdJetBrains/intellij-obsolete-plugins/aidebugger/CLAUDE.md
Changelog is maintained manually in CHANGELOG.md. Add entries under ## [Unreleased] for notable user-visible changes. On release, rename [Unreleased] to the version with date. No tooling automation — update by hand. This plugin uses the monorepo Bazel/IML build infrastructure (not Gradle). IML files define module dependencies. Run ./build/jpsModelToBazel.cmd from the repo root after modifying IML files to regenerate BUILD.bazel files.
JetBrainsCLAUDE.mdJetBrains/intellij-obsolete-plugins/ai-playground/CLAUDE.md
No summary in the file. Open it to read it.
BlockReads credentialsCLAUDE.mdblock/ghost/CLAUDE.md
Agents now do the making — screens, emails, empty states, sentences. What they cannot reliably preserve is the brand behind that work: the stance, the density, the restraint, the trust moves, and the choices that make an output feel intentional. An agent holds nothing it isn't handed. ghost hands it the brand as a portable steering packet: a repo-local .ghost/ package, a flat corpus of prose nodes read before anything is made. The public npm shape is one package, @design-intelligence/ghost,…
MicrosoftCLAUDE.mdmicrosoft/Agent365-python/CLAUDE.md
Microsoft Agent 365 SDK for Python - Enterprise-grade extensions for building production-ready AI agents for M365, Teams, Copilot Studio, and Webchat. This is a uv workspace with eight packages for Notifications, Runtime, and Tooling. Observability is retired from this repository. Do not reintroduce its packages, sources, imports, or release dependencies. Use the separately maintained microsoft-opentelemetry distro; see migration and owner actions. Python Version: 3.11+ (3.11 and 3.12 tested in CI) For detailed architecture and design documentation, see docs/design.md, which includes:…
SentryCLAUDE.mdgetsentry/XcodeBuildMCP-iOS-Template/template/CLAUDE.md
This is a native iOS application built with Swift 6.1+ and SwiftUI. The codebase targets iOS 18.0 and later, allowing full use of modern Swift and iOS APIs. All concurrency is handled with Swift Concurrency (async/await, actors, @MainActor isolation) ensuring thread-safe code. The project follows a workspace + SPM package architecture: Important: All development work should be done in the YourAppPackage Swift Package, not in the app project. The app project is merely a thin wrapper that imports and launches…
MicrosoftReads credentialsCLAUDE.mdmicrosoft/power-platform-playwright-samples/CLAUDE.md
This file is the entry point for customers and AI assistants working with this repo. It walks through install → run → debug → generate-new-tests, plus reference material that AI…
DatabricksCLAUDE.mddatabricks/databricks-jdbc/CLAUDE.md
When adding or changing an error emitted by the JDBC driver: - Use DatabricksDriverErrorCode wherever a driver-owned error code applies. Reuse an existing code when its meaning matches; otherwise add a uniquely numbered enum value. - Add or update a test that verifies the emitted error name and numeric code. - Record whether the error is a driver, server, or user error in the maintainers' OSS JDBC telemetry dashboard and automation taxonomy. Link that update in the PR, or explicitly…
Microsoft AzureCLAUDE.mdAzure/aks-app-routing-operator/CLAUDE.md
The AKS Web Application Routing Operator is a Kubernetes operator that manages resources for AKS Application Routing functionality. It's built using the Kubebuilder framework and manages several Custom Resource Definitions (CRDs) for routing, DNS, and certificate management. Some test packages (e.g. pkg/controller/dns/) use controller-runtime's envtest and require Kubernetes API server binaries. Use setup-envtest to install them and set KUBEBUILDER_ASSETS: If setup-envtest use -p path doesn't expand correctly in your shell, run setup-envtest use first to see the path, then set…
Microsoft AzureCLAUDE.mdAzure/aks-app-routing-operator/pkg/manifests/CLAUDE.md
This package stores Kubernetes manifests that are served to customers. These manifests are used to deploy various components of the App Routing Operator. The test fixtures in this package are generated from the Helm charts. To regenerate the fixtures, run the tests with the GENERATE_FIXTURES environment variable set to true:
Microsoft AzureCLAUDE.mdAzure/ARO-HCP/CLAUDE.md
This is the main repository for Red Hat OpenShift on Azure (ARO) using the Hosted Control Planes (HCP) architecture. It contains some of the code for the required microservices along with most of the configuration and pipeline definition necessary to deploy it. For integration tests (require Cosmos emulator): Rendered config changes under config/rendered/ must be committed with your PR. Run make verify-materialize to check. Note: make test runs envtest-setup automatically, but if you run go test directly in a module…
Microsoft AzureCLAUDE.mdAzure/ARO-HCP/config/CLAUDE.md
Docs on how these configs work: - docs/configuration.md - docs/bicep.md if you're editing bicep files When making changes to configs/config*.yaml you might need to update the appropriate schema files. After updating configs, verify and render new config by running: - cd config; make materialize These might result in changes to rendered configs which must be attached to any PRs. Examples: - Commit 4a276befedc7f8a33692c6d8d6746e22344f195c adds a new bool parameter set to True by default, but False in 'dev' cloud envs. Commit…
A file Claude Code reads at the start of every session. It holds the commands, conventions and warnings the agent needs for this project.
At the repository root. Claude Code also reads CLAUDE.md files in subdirectories when it works there.
Build and test commands, the project's layout, conventions that aren't obvious from the code, and mistakes to avoid. Short files tend to work better than long ones.
Claude Code reads CLAUDE.md; most other agents read AGENTS.md. Many projects keep one and point the other at it.
