Cloudflare Deployment Assistant
Guides Cloudflare product selection and deployment for applications and infrastructure.
The documentation requires checking Wrangler authentication before deployment and shows API-token, environment-variable, and secret-management patterns, providing partial authentication and data-flow awareness. Points are deducted because deployment has external effects but explicit user confirmation, least-privilege configuration, change scoping, rollback, and post-deployment verification are not specified; the escalation advice also lacks detailed isolation and risk controls.
Authentication guidance, product decision trees, configuration examples, and troubleshooting cover many plausible paths. Points are deducted because the static material contains no execution or test evidence, many referenced directories are not included in the supplied material, and some examples appear questionable, such as using await inside a non-async onStart; abnormal-input, deployment-failure, and version-compatibility feedback is thin. Static calibration limits this to 10 or below.
The trigger is clear, and the decision trees map compute, storage, AI, networking, and deployment needs to Cloudflare products. Points are deducted because non-fit scenarios, required inputs, output contracts, and trigger boundaries are underdefined; Chinese-language support and mainland-China reachability for Cloudflare, npm, OAuth, and API services are not addressed.
The main document uses progressive sections for prerequisites, decision trees, product indexes, and troubleshooting, with generally clear naming and references. Points are deducted because the supplied license metadata is unknown, and no versioning, changelog, explicit maintenance owner, or update path is shown; the repository README also states that the repository is deprecated, while dependency and complete-reference installation notes are incomplete.
The skill can provide Cloudflare product selection, authentication, configuration, and deployment commands, so it plausibly supports basic deployment preparation. Points are deducted because its scope is very broad, key implementation details depend on many unprovided reference files, and no complete end-to-end output, returned URL, or configuration validation is statically verified; substantial manual checking may remain. Static calibration limits this to 7 or below.
The supplied files contain inspectable commands, configuration snippets, code examples, and troubleshooting notes, giving limited auditability. Points are deducted because there is no test suite, CI execution record, third-party reproduction evidence, or corroborating evidence from multiple sources; many claims about behavior and compatibility cannot be independently verified from the supplied files. Static calibration limits this to 5 or below.
- Deployment commands create external infrastructure changes; add per-operation user confirmation, least-privilege tokens, change previews, and a verifiable rollback process before use.
- Do not place API tokens, OAuth credentials, or provider keys from the examples in source code; verify reachability of Cloudflare, npm, OAuth, and related APIs in the target network.
- The repository is marked deprecated, and the supplied material lacks version pinning, test execution evidence, and complete reference files; verify current official documentation and dependency compatibility before use.
What it does & when to use it
This skill supports deployment to the Cloudflare platform, including Workers, Pages, and related services. Decision trees map compute, storage, AI, networking, security, media, and infrastructure-as-code needs to Cloudflare products. It requires Wrangler authentication before deployment and supports local OAuth login or a CI/CD API token. The repository is deprecated, so current installation and plugin examples should be taken from the OpenAI Plugins repository.
Maps deployment, hosting, publishing, and Cloudflare setup requests to relevant products; loads detailed material from references/; runs npx wrangler whoami before wrangler deploy, wrangler pages deploy, or npm run deploy; and advises retrying with elevated network access when sandbox networking blocks deployment.
- A developer deploying edge serverless functions to Cloudflare Workers.
- A team publishing a full-stack web application to Cloudflare Pages through Git deployments.
- An engineer choosing among KV, D1, R2, Queues, and Durable Objects for application data.
- A developer selecting Workers AI, Vectorize, or the Agents SDK for a Cloudflare AI workload.
- An operations engineer managing Cloudflare infrastructure through Terraform, Pulumi, or the REST API.
Pros & cons
- Covers many Cloudflare product areas with need-based decision trees.
- Specifies an authentication check before deployment.
- Supports both local OAuth and CI/CD API-token workflows.
- Includes guidance for network-related deployment failures.
- The repository is explicitly deprecated and should not be the sole source for current examples.
- The supplied material mainly provides routing and operational guidance; detailed product references are not included here.
- The source provides no test-suite, platform-coverage, or deployment-success evidence.
- Deployment may require elevated network access and authenticated Cloudflare account access.
How to install
The skill is located at skills/.curated/cloudflare-deploy. The README gives this curated-skill installation command: $skill-installer cloudflare-deploy. Restart Codex after installation. The provided sources do not document additional environment setup.
How to use
Give the agent a request such as “Deploy this project to Cloudflare Workers” or “Publish this website to Cloudflare Pages.” Before deployment, run npx wrangler whoami; use wrangler login for local OAuth authentication or set CLOUDFLARE_API_TOKEN in CI/CD. The supplied sources do not provide complete product-specific deployment commands or configuration examples.