coding-standards

$npx mdskill add cloudflare/sandbox-sdk/coding-standards

Enforce TypeScript rules and style for code quality.

  • Prevents unsafe type usage and enforces naming conventions.
  • Depends on project type definitions and style guides.
  • Identifies violations of type safety and acronym rules.
  • Reports specific rule breaches with actionable corrections.

SKILL.md

.github/skills/coding-standardsView on GitHub ↗
---
name: coding-standards
description: Use when writing or reviewing TypeScript in this repo. Covers the no-`any` rule and where to put new types, the uppercase-acronym style guide, and the rules for code comments (no historical context). (project)
---

# Coding Standards

## TypeScript: No `any`

**Never use `any`** unless absolutely necessary — and that should be a final resort.

Process when you reach for `any`:

1. Look for an existing type that fits. Most domains already have one.
2. If no suitable type exists, define a proper one in the right location:
   - **Shared types** → `packages/shared/src/types.ts` or relevant subdirectory
   - **SDK-specific types** → `packages/sandbox/src/clients/types.ts` or the appropriate client file
   - **Container-specific types** → under `packages/sandbox-container/src/` with appropriate naming
3. Use the new type everywhere it applies — don't leave one-off shapes scattered around.

This catches errors at compile time instead of runtime and keeps the codebase consistent.

## Style: Uppercase Acronyms

When an acronym appears inside a camelCase or PascalCase identifier, keep it **fully uppercase**:

| ✅ Do             | ❌ Don't          |
| ----------------- | ----------------- |
| `SandboxRPCAPI`   | `SandboxRpcApi`   |
| `containerURL`    | `containerUrl`    |
| `parseHTTPHeader` | `parseHttpHeader` |
| `getAPIKey`       | `getApiKey`       |

Applies to all acronyms: API, URL, HTTP, RPC, SSE, SSH, DNS, ID, etc.

**Exception:** library-provided names keep their original casing (e.g. capnweb's `RpcTarget` stays `RpcTarget`).

## Code Comments

**Write comments for future readers, not for the current conversation.**

Comments should describe the current state of the code. A developer reading the code months later won't have context about bugs that were fixed, conversations that happened, or earlier implementations.

### Don't reference historical context

```typescript
// ❌ Bad: references a bug the reader knows nothing about
// Uses character tracking to avoid the bug where indexOf('') returns wrong position

// ❌ Bad: implies something was wrong before
// Start the server with proper WebSocket typing

// ❌ Bad: "prevent" implies there was a problem to prevent
// Assign synchronously to prevent race conditions
```

### Do describe current behavior and design intent

```typescript
// ✅ Good: describes what the code does now
// Returns parsed events and any remaining unparsed content

// ✅ Good: explains design rationale without historical context
// Assigned synchronously so concurrent callers share the same connection attempt

// ✅ Good: explains a non-obvious implementation choice
// Uses IIFE to ensure promise exists before any await points
```

### Smell test

If your comment contains "to avoid", "to fix", "to prevent", "instead of", or "properly" — reconsider whether you're describing current behavior or quietly referencing something that no longer exists. Rewrite to describe what the code does now and why this design was chosen.

## API Design

When adding or modifying SDK methods:

- Use clear, descriptive names that indicate what the method does
- Validate inputs before passing to container APIs
- Provide helpful error messages with context (use the custom error classes in `packages/shared/src/errors/`)

More from cloudflare/sandbox-sdk

SkillDescription
architectureUse when navigating the codebase for the first time, adding a new client method, adding a new container handler/service, or understanding how a request flows from Worker through the Sandbox DO into the container. Covers the three-layer architecture, client pattern, container runtime structure, and monorepo layout. (project)
changesetsUse when creating a changeset, preparing a release, or bumping versions. Covers which packages to reference, how to write user-facing changeset descriptions, the release automation flow, and the npm/Docker version sync requirement. (project)
examplesUse when working in the examples/ directory, running an example with wrangler dev, adding a new example, or answering questions about EXPOSE directives and the local Docker dev loop. (project)
git-commitUse when creating git commits to ensure commit messages follow project standards. Applies the 7 rules for great commit messages with focus on conciseness and imperative mood.
loggingUse when adding logs, debugging, or working with the Logger across the SDK and container runtime. Covers the constructor-injection pattern, child loggers, env-var configuration, and test mocking. (project)
sandbox-bridgeUse when you need to exercise a real, running Sandbox deployment via HTTP — for example to validate SDK changes against a live container, reproduce a user-reported issue, or experiment with the API (including FUSE bucket mounts) without spinning up `wrangler dev`. Documents the Sandbox bridge worker reachable via `SANDBOX_WORKER_URL` + `SANDBOX_API_KEY` when the host injects them.
session-executionUse when working on or reviewing session execution, command handling, shell state, FIFO-based streaming, or stdout/stderr separation. Relevant for session.ts, command handlers, exec/execStream, or anything involving shell process management. (project)
testingUse when writing or running tests for this project. Covers unit vs E2E test decisions, test file locations, mock patterns, and project-specific testing conventions. (project)