mcpbeat

Workers Best Practices

openai/workers-best-practices

Reviews and authors Cloudflare Workers code against production best practices. Load when writing new Workers, reviewing Worker code, configuring wrangler.jsonc, or checking for common Workers anti-patterns (streaming, floating promises, global state, secrets, bindings, observability). Biases towards retrieval from Cloudflare docs over pre-trained knowledge.

8k tokens
context cost
the whole folder, loaded on every use
4
files
instructions only
1
copies elsewhere
how many repositories repackaged it
4915
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/openai/plugins --skill workers-best-practices

What comes with it

24 750 bytes besides the instruction
agents/openai.yaml
references/review.md
references/rules.md

The instruction itself

14 sections, as written by the author

Your knowledge of Cloudflare Workers APIs, types, and configuration may be outdated. Prefer retrieval over pre-training for any Workers code task — writing or reviewing.

Retrieval Sources

Fetch the latest versions before writing or reviewing Workers code. Do not rely on baked-in knowledge for API signatures, config fields, or binding shapes.

| Source | How to retrieve | Use for |

|--------|----------------|---------|

| Workers best practices | Fetch https://developers.cloudflare.com/workers/best-practices/workers-best-practices/ | Canonical rules, patterns, anti-patterns |

| Workers types | See references/review.md for retrieval steps | API signatures, handler types, binding types |

| Wrangler config schema | node_modules/wrangler/config-schema.json | Config fields, binding shapes, allowed values |

| Cloudflare docs | Search tool or https://developers.cloudflare.com/workers/ | API reference, compatibility dates/flags |

FIRST: Fetch Latest References

Before reviewing or writing Workers code, retrieve the current best practices page and relevant type definitions. If the project's node_modules has an older version, prefer the latest published version.

# Fetch latest workers types
mkdir -p /tmp/workers-types-latest && \
  npm pack @cloudflare/workers-types --pack-destination /tmp/workers-types-latest && \
  tar -xzf /tmp/workers-types-latest/cloudflare-workers-types-*.tgz -C /tmp/workers-types-latest
# Types at /tmp/workers-types-latest/package/index.d.ts

Reference Documentation

  • references/rules.md — all best practice rules with code examples and anti-patterns
  • references/review.md — type validation, config validation, binding access patterns, review process

Rules Quick Reference

Configuration

| Rule | Summary |

|------|---------|

| Compatibility date | Set compatibility_date to today on new projects; update periodically on existing ones |

| nodejs_compat | Enable the nodejs_compat flag — many libraries depend on Node.js built-ins |

| wrangler types | Run wrangler types to generate Env — never hand-write binding interfaces |

| Secrets | Use wrangler secret put, never hardcode secrets in config or source |

| wrangler.jsonc | Use JSONC config for non-secret settings — newer features are JSON-only |

Request & Response Handling

| Rule | Summary |

|------|---------|

| Streaming | Stream large/unknown payloads — never await response.text() on unbounded data |

| waitUntil | Use ctx.waitUntil() for post-response work; do not destructure ctx |

Architecture

| Rule | Summary |

|------|---------|

| Bindings over REST | Use in-process bindings (KV, R2, D1, Queues) — not the Cloudflare REST API |

| Queues & Workflows | Move async/background work off the critical path |

| Service bindings | Use service bindings for Worker-to-Worker calls — not public HTTP |

| Hyperdrive | Always use Hyperdrive for external PostgreSQL/MySQL connections |

Observability

| Rule | Summary |

|------|---------|

| Logs & Traces | Enable observability in config with head_sampling_rate; use structured JSON logging |

Code Patterns

| Rule | Summary |

|------|---------|

| No global request state | Never store request-scoped data in module-level variables |

| Floating promises | Every Promise must be awaited, returned, voided, or passed to ctx.waitUntil() |

Security

| Rule | Summary |

|------|---------|

| Web Crypto | Use crypto.randomUUID() / crypto.getRandomValues() — never Math.random() for security |

| No passThroughOnException | Use explicit try/catch with structured error responses |

Anti-Patterns to Flag

| Anti-pattern | Why it matters |

|-------------|----------------|

| await response.text() on unbounded data | Memory exhaustion — 128 MB limit |

| Hardcoded secrets in source or config | Credential leak via version control |

| Math.random() for tokens/IDs | Predictable, not cryptographically secure |

| Bare fetch() without await or waitUntil | Floating promise — dropped result, swallowed error |

| Module-level mutable variables for request state | Cross-request data leaks, stale state, I/O errors |

| Cloudflare REST API from inside a Worker | Unnecessary network hop, auth overhead, added latency |

| ctx.passThroughOnException() as error handling | Hides bugs, makes debugging impossible |

| Hand-written Env interface | Drifts from actual wrangler config bindings |

| Direct string comparison for secret values | Timing side-channel — use crypto.subtle.timingSafeEqual |

| Destructuring ctx (const { waitUntil } = ctx) | Loses this binding — throws "Illegal invocation" at runtime |

| any on Env or handler params | Defeats type safety for all binding access |

| as unknown as T double-cast | Hides real type incompatibilities — fix the design |

| implements on platform base classes (instead of extends) | Legacy — loses this.ctx, this.env. Applies to DurableObject, WorkerEntrypoint, Workflow |

| env.X inside platform base class | Should be this.env.X in classes extending DurableObject, WorkerEntrypoint, etc. |

Review Workflow

  • Retrieve — fetch latest best practices page, workers types, and wrangler schema
  • Read full files — not just diffs; context matters for binding access patterns
  • Check types — binding access, handler signatures, no any, no unsafe casts (see references/review.md)
  • Check config — compatibility_date, nodejs_compat, observability, secrets, binding-code consistency
  • Check patterns — streaming, floating promises, global state, serialization boundaries
  • Check security — crypto usage, secret handling, timing-safe comparisons, error handling
  • Validate with toolsnpx tsc --noEmit, lint for no-floating-promises
  • Reference rules — see references/rules.md for each rule's correct pattern

Scope

This skill covers Workers-specific best practices and code review. For related topics:

  • Durable Objects: load the durable-objects skill
  • Workflows: see Rules of Workflows
  • Wrangler CLI commands: load the wrangler skill

Principles

  • Be certain. Retrieve before flagging. If unsure about an API, config field, or pattern, fetch the docs first.
  • Provide evidence. Reference line numbers, tool output, or docs links.
  • Focus on what developers will copy. Workers code in examples and docs gets pasted into production.
  • Correctness over completeness. A concise example that works beats a comprehensive one with errors.

Repackaged in 1 other repositories

same content, different owner
cloudflare/skills open on GitHub →

How to use it

Copy the folder

Take openai/workers-best-practices 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.

Install what it needs

The instructions reference npx. Without those the skill loads but fails at the first command.