mcpbeat Sign in

Changesets Agent Skill

Use 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)

1k tokens
context cost
the whole folder, loaded on every use
1
files
instructions only
0
copies elsewhere
how many repositories repackaged it
1089
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/cloudflare/sandbox-sdk --skill changesets

The instruction itself

7 sections, as written by the author

Changesets & Releases

This repository uses changesets to create Version Packages PRs and maintain changelog/version files. Stable and prerelease artifact publishing are handled by the release orchestrator. Create a changeset whenever your change affects published packages.

Creating a Changeset

A changeset should be created when there is a change that is observable to a consumer of the

@cloudflare/sandbox package. This includes:

  • Changes to the API surface area, new methods, deprecations or removals.
  • Changes to the performance or security characteristics of the SDK.
  • Bug fixes that are user visible.

Create a new file in .changeset/ (e.g. .changeset/your-feature-name.md):

---
'@cloudflare/sandbox': patch
---

Brief description of your change

Rules

Only reference @cloudflare/sandbox. Never list @repo/shared or @repo/sandbox-container — those are internal workspace packages and must not be versioned independently. Changes to them flow through the public package. Pre-commit hooks and CI enforce this.

Use patch for almost everything. The SDK is in beta:

  • patch — all normal changes (features, fixes, refactors)
  • minor — breaking changes only
  • major — never

Writing the Description

Important: Changeset files should only reference @cloudflare/sandbox, never @repo/shared or @repo/sandbox-container. These internal packages should not be versioned independently - changes to them flow through the public package. Pre-commit hooks and CI will validate this rule.

Important: Write for end users. Changeset descriptions appear in GitHub releases - they're user-facing documentation, not internal notes.

  • Focus on the problem solved and the benefit, not technical implementation details.
  • Keep it short. Each changeset entry should aim to be a couple of sentences, no more than a single paragraph.
  • Include a code example showing how to enable or use the feature when applicable.
# Bad - technical/internal focused

Add WebSocket transport for request multiplexing over a single connection

# Good - user-focused with clear benefit and usage

Add WebSocket transport to avoid sub-request limits in Workers and Durable Objects.
Enable with `useWebSocket: true` in sandbox options.

Release Automation

Releases run via .github/workflows/release.yml. There is no manual publishing step.

  • Merge a PR that contains a changeset.
  • The Changesets action opens (or updates) a "Version Packages" PR.
  • Merging the Version Packages PR triggers the stable release workflow.
  • The release orchestrator publishes and verifies the npm package, Docker Hub images, CF Registry public library images, GitHub Release, and standalone binary assets.
  • Prerelease workflows also publish through the release orchestrator, which verifies the npm dist-tag, Docker Hub images, CF Registry public library images, and optional moving Docker aliases before succeeding.
  • Release workflow reruns converge missing artifacts, so Docker images, GitHub releases, binaries, and prerelease aliases are not gated on npm being newly published in the current attempt.

Version Synchronization

Docker image version MUST match the npm package version. This is enforced via ARG SANDBOX_VERSION in packages/sandbox/Dockerfile. Don't try to release them out of band.

  • SDK version is tracked in packages/sandbox/src/version.ts
  • Images build for linux/amd64 only, matching Cloudflare's production runtime (ARM Macs use Rosetta/QEMU locally, preserving dev/prod parity)
  • Images publish to both Docker Hub and registry.cloudflare.com/library/sandbox:{version} (with -python, -opencode, -musl variants). Any authenticated Cloudflare customer can pull from the library/ namespace without our account ID.

Checklist

  • [ ] Filename in .changeset/ is descriptive (fix-stream-encoding.md, not patch.md)
  • [ ] Only @cloudflare/sandbox is listed
  • [ ] Bump type is patch (or minor for breaking changes)
  • [ ] Description explains the user-visible problem and benefit
  • [ ] Usage hint is included when relevant (flag name, option, method)

Other skills for the same job

different authors, same section of the catalogue
Modal
by christophacham
×3

Run Python code in the cloud with serverless containers, GPUs, and autoscaling. Use when deploying ML models, running batch processing jobs, scheduling compute-intensive tasks, or serving APIs that require GPU acceleration or dynamic scaling.

17k tokens
Github Workflow Automation
by ComeOnOliver
×3

Advanced GitHub Actions workflow automation with AI swarm coordination, intelligent CI/CD pipelines, and comprehensive repository management

9k tokens
Gcloud
by Dicklesworthstone
×2

Google Cloud Platform CLI - manage GCP resources including Compute Engine, Cloud Run, GKE, Cloud Functions, Storage, BigQuery, and more.

2k tokens
Backend Architect
by ComeOnOliver
×2

Expert backend architect specializing in scalable API design, microservices architecture, and distributed systems. Masters REST/GraphQL/gRPC APIs, event-driven architectures, service mesh patterns, and modern backend frameworks. Handles service boundary definition, inter-service communication, resilience patterns, and observability. Use PROACTIVELY when creating new backend services or APIs.

7k tokens
Modal
by ComeOnOliver
×2

Run Python code in the cloud with serverless containers, GPUs, and autoscaling. Use when deploying ML models, running batch processing jobs, scheduling compute-intensive tasks, or serving APIs that require GPU acceleration or dynamic scaling.

37k tokens
Aspire
by github
vendor ×1

Aspire skill covering the Aspire CLI, AppHost orchestration, service discovery, integrations, MCP server, VS Code extension, Dev Containers, GitHub Codespaces, templates, dashboard, and deployment. Use when the user asks to create, run, debug, configure, deploy, or troubleshoot an Aspire distributed application.

21k tokens
Bigquery Pipeline Audit
by github
vendor ×1

Audits Python + BigQuery pipelines for cost safety, idempotency, and production readiness. Returns a structured report with exact patch locations.

1k tokens
Msstore CLI
by github
vendor ×1

Microsoft Store Developer CLI (msstore) for publishing Windows applications to the Microsoft Store. Use when asked to configure Store credentials, list Store apps, check submission status, publish submissions, manage package flights, set up CI/CD for Store publishing, or integrate with Partner Center. Supports Windows App SDK/WinUI, UWP, .NET MAUI, Flutter, Electron, React Native, and PWA applications.

4k tokens

How to use it

Copy the folder

Take cloudflare/changesets from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.