agentleFS
ckorhonenckorhonen/claude-skills/skills/cloudflare-manager/SKILL.md

cloudflare-manager

Skill17 starsChanged 4 months agoMITView on GitHub

Comprehensive Cloudflare account management for deploying Workers, KV Storage, R2, Pages, DNS, and Routes. Use when deploying cloudflare services, managing worker containers, configuring KV/R2 storage, or setting up DNS/routing. Requires CLOUDFLARE_API_KEY in .env and Bun runtime with dependencies installed.

  • Reads credentials

The file

Show its rules and text

We split this skill into the rules it gives an agent

50 rules. About 0 are common, 27 are distinctive, and 23 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈27

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 23

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    Before using this skill for the first time:

    line 12
  2. Very specific:

    Create a .env file in your project root:

    line 22
  3. Distinctive:

    Use "Edit Cloudflare Workers" template (or create custom token)

    line 31
  4. Distinctive:

    Run validation to verify your API key and check permissions:

    line 41
  5. Very specific:

    If validation fails with 401/403: Check your API token is correct in .env

    line 60
  6. Distinctive:

    If validation fails with network error: Check internet connection

    line 61
  7. Very specific:

    Use --no-cache flag to force fresh validation: bun scripts/validate-api-key.ts --no-cache

    line 62
  8. Very specific:

    Run bun scripts/validate-api-key.ts to populate this section with your current permissions.

    line 67
  9. Distinctive:

    Create KV namespace for session caching:

    line 107
  10. Distinctive:

    Create R2 bucket for user uploads:

    line 113
  11. Distinctive:

    Create Pages project (if doesn't exist):

    line 152
  12. Distinctive:

    When creating a token at https://dash.cloudflare.com/profile/api-tokens, use the "Edit Cloudflare Workers" template — don't use "Read All" …

    line 220
  13. Very specific:

    Before running wrangler login, kill all other Wrangler processes:

    line 247
  14. Very specific:

    Only start wrangler dev or Pages after login completes

    line 255
  15. Distinctive:

    Worker scripts: Keep under 1MB (smaller = faster cold start)

    line 266
  16. Very specific:

    Use .cloudflare/ ignore patterns to exclude node_modules, build caches, etc.

    line 269
  17. Very specific:

    If .env file is missing or CLOUDFLARE_API_KEY is not set:

    line 428
  18. Distinctive:

    If API token lacks required permissions:

    line 438
  19. Distinctive:

    If too many requests are made:

    line 451
  20. Distinctive:

    If API is unreachable:

    line 460
  21. Very specific:

    Never commit .env files - always add to .gitignore

    line 470
  22. Distinctive:

    Use token-based authentication (not API keys)

    line 471
  23. Distinctive:

    Use least-privilege principle: only grant required permissions

    line 473
  24. Distinctive:

    Use KV storage for frequently-read data (not frequently-written)

    line 478
  25. Distinctive:

    Use R2 for large files (KV has 25MB limit per key)

    line 479
  26. Distinctive:

    Keep worker scripts under 1MB for faster cold starts

    line 481
  27. Very specific:

    Test locally first: wrangler dev for local testing

    line 484
  28. Distinctive:

    Use staging environment before production

    line 485
  29. Distinctive:

    Delete unused workers, namespaces, and buckets

    line 497
  30. Very specific:

    All scripts are located in ~/.claude/skills/cloudflare-manager/scripts/:

    line 516
  31. Very specific:

    kv-storage.ts: Create and manage KV namespaces and key-value pairs

    line 520
  32. Very specific:

    r2-storage.ts: Create and manage R2 buckets and objects

    line 521
  33. Very specific:

    Check script syntax: node --check ./worker.js

    line 572
  34. Very specific:

    Verify file exists: ls -lh ./worker.js

    line 573
  35. Distinctive:

    Check worker name is valid (alphanumeric, hyphens, underscores only)

    line 575
  36. Distinctive:

    Verify you're using namespace ID (not name) in commands

    line 585
  37. Distinctive:

    Check namespace wasn't deleted

    line 586
  38. Distinctive:

    Ensure API token has KV Storage permissions

    line 587
  39. Very specific:

    Use existing bucket: bun scripts/r2-storage.ts list-buckets

    line 598
  40. Very specific:

    Check deployment status: bun scripts/pages.ts list-deployments project-name

    line 608
  41. Very specific:

    Verify zone exists: bun scripts/dns-routes.ts list-zones

    line 620
  42. Distinctive:

    Ensure domain is added to Cloudflare and active

    line 621
  43. Very specific:

    Check nameservers point to Cloudflare: dig NS yourdomain.com

    line 622
  44. Distinctive:

    Verify API token has Zone > DNS > Edit permission

    line 623
  45. Distinctive:

    Wait 1-2 minutes before retrying manually

    line 633
  46. Very specific:

    Create .env file in project root (not skill directory)

    line 644
  47. Very specific:

    Verify file content: cat .env | grep CLOUDFLARE_API_KEY

    line 645
  48. Very specific:

    Ensure no extra spaces: CLOUDFLARE_API_KEY=token (no spaces around =)

    line 646
  49. Very specific:

    Run commands from project root where .env exists

    line 647
  50. Distinctive:

    All API requests use HTTPS

    line 659

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Cloudflare Manager

Comprehensive Cloudflare service management skill that enables deployment and configuration of Workers, KV Storage, R2 buckets, Pages, DNS records, and routing. Automatically validates API credentials, extracts deployment URLs, and provides actionable error messages.

Initial Setup

Before using this skill for the first time:

  1. Install Dependencies
    cd ~/.claude/skills/cloudflare-manager
    bun install
  2. Configure API Key

    Create a .env file in your project root:

    CLOUDFLARE_API_KEY=your_api_token_here
    CLOUDFLARE_ACCOUNT_ID=your_account_id  # Optional, auto-detected

    Getting your API token:

    • Visit https://dash.cloudflare.com/profile/api-tokens
    • Click "Create Token"
    • Use "Edit Cloudflare Workers" template (or create custom token)
    • Required permissions:
      • Account > Workers Scripts > Edit
      • Account > Workers KV Storage > Edit
      • Account > Workers R2 Storage > Edit
      • Account > Cloudflare Pages > Edit
      • Zone > DNS > Edit (if using custom domains)
  3. Validate Credentials

    Run validation to verify your API key and check permissions:

    cd ~/.claude/skills/cloudflare-manager
    bun scripts/validate-api-key.ts

    Expected output:

    ✅ API key is valid!
    ℹ️  Token Status: active
    ℹ️  Account: Your Account Name (abc123...)
    
    🔑 Granted Permissions:
      ✅ Workers Scripts: Edit
      ✅ Workers KV Storage: Edit
      ✅ Workers R2 Storage: Edit

    Troubleshooting validation:

    • If validation fails with 401/403: Check your API token is correct in .env
    • If validation fails with network error: Check internet connection
    • Use --no-cache flag to force fresh validation: bun scripts/validate-api-key.ts --no-cache

Current API Permissions

<!-- PERMISSIONS_START -->

Run bun scripts/validate-api-key.ts to populate this section with your current permissions.

<!-- PERMISSIONS_END -->

Examples

Example 1: Deploy New Worker API

User request:

"Set up a Cloudflare worker for handling API requests and give me the URL"

Workflow:

  1. Validate API credentials exist in .env
  2. Deploy worker script:
    bun scripts/workers.ts deploy api-handler ./worker.js
  3. Extract Cloudflare-generated URL from response
  4. Return URL: https://api-handler.username.workers.dev

Expected output:

✅ Worker deployed successfully!
📍 URL: https://api-handler.username.workers.dev
🆔 Worker ID: abc123def456

Time: 2-5 seconds


Example 2: Full Application Stack (Worker + KV + R2)

User request:

"Set up a complete Cloudflare application with worker, key-value storage, and file storage"

Workflow:

  1. Create KV namespace for session caching:
    bun scripts/kv-storage.ts create-namespace app-sessions
    # Returns: namespace-id-123
  2. Create R2 bucket for user uploads:
    bun scripts/r2-storage.ts create-bucket app-uploads
    # Returns: Bucket created: app-uploads
  3. Deploy worker with bindings:
    bun scripts/workers.ts deploy app-api ./worker.js \
      --kv-binding SESSION_STORE=namespace-id-123 \
      --r2-binding UPLOADS=app-uploads
  4. Configure custom domain route:
    bun scripts/dns-routes.ts create-route example.com \
      "api.example.com/*" app-api

Expected output:

✅ KV namespace created: app-sessions (namespace-id-123)
✅ R2 bucket created: app-uploads
✅ Worker deployed: https://app-api.username.workers.dev
✅ Route configured: api.example.com/* → app-api

Time: 8-12 seconds total


Example 3: Deploy Static Site to Pages

User request:

"Deploy my static site to Cloudflare Pages"

Workflow:

  1. Create Pages project (if doesn't exist):
    bun scripts/pages.ts deploy my-site ./dist
  2. For first deployment with files, use Wrangler:
    npx wrangler pages deploy ./dist --project-name=my-site
  3. Set environment variables:
    bun scripts/pages.ts set-env my-site API_URL https://api.example.com
    bun scripts/pages.ts set-env my-site DEBUG false --env production

Expected output:

✅ Pages project created: my-site
📍 URL: https://my-site.pages.dev
🚀 Deploying files... (via Wrangler)
✅ Deployment complete!
✅ Environment variables set

Time: 15-30 seconds (depends on site size)


Quick Start Guide

Deploy a Worker Container

To deploy a new worker container sandbox:

# Using the skill
bun scripts/workers.ts deploy worker-name ./worker-script.js

What happens:

  • Creates new worker container
  • Deploys JavaScript/TypeScript code
  • Automatically extracts and returns Cloudflare-generated URL (e.g., https://worker-name.username.workers.dev)
  • Returns worker ID and configuration

Example conversation:

User: "Set up and deploy a new cloudflare worker container sandbox named 'api-handler' and return the URL"
Claude: [Deploys worker using bun scripts/workers.ts deploy api-handler ./worker.js]
       Returns URL: https://api-handler.username.workers.dev

Exit codes:

  • 0: Success - worker deployed and URL returned
  • 1: Failure - check error message for details

Performance: Deployment typically completes in 2-5 seconds

Common Pitfalls

⚠️ Most Critical Issues — Address these first if you encounter problems:

1. API Token Scoping (Most Common Cause of Failures)

The Problem: API tokens with insufficient or incorrectly scoped permissions fail silently with 403 errors, making debugging very difficult.

How to Avoid:

  • When creating a token at https://dash.cloudflare.com/profile/api-tokens, use the "Edit Cloudflare Workers" template — don't use "Read All" or custom scopes without the right permissions
  • Required minimum scopes:
    • Account > Workers Scripts > Edit
    • Account > Workers KV Storage > Edit
    • Account > Workers R2 Storage > Edit
    • Zone > DNS > Edit (only if using custom domains)
  • Verify your token immediately after creation:
    bun scripts/validate-api-key.ts
    This shows exactly which permissions you have. Do this every time you create or update a token.

Quick Diagnosis:

# If you see "Insufficient permissions", stop and fix the token:
bun scripts/validate-api-key.ts
# If any required permission shows ❌, update token at:
# https://dash.cloudflare.com/profile/api-tokens

2. Wrangler OAuth Port Conflicts

The Problem: wrangler login uses port 8976 for the OAuth callback. If another wrangler dev or wrangler pages dev process is already running, the login flow will fail silently — the browser tab opens but never receives the OAuth redirect.

How to Avoid:

  • Before running wrangler login, kill all other Wrangler processes:
    pkill -f wrangler
  • Then run login:
    wrangler login
  • Only start wrangler dev or Pages after login completes

Why This Happens: Wrangler doesn't queue OAuth callbacks — the first process to bind port 8976 gets the callback. If that's a dev server, login attempts silently fail.


3. Deployment Timeouts (Large Projects)

The Problem: Workers and Pages deployments timeout if the project is too large or network connectivity is unstable. Timeouts also occur during peak Cloudflare load.

How to Avoid:

  • Worker scripts: Keep under 1MB (smaller = faster cold start)
  • Pages deployments:
    • Projects with >1000 files may take 5-10 minutes
    • Use .cloudflare/ ignore patterns to exclude node_modules, build caches, etc.
    • For large deployments, use Wrangler CLI directly:
      npx wrangler pages deploy ./dist --project-name=my-app
  • Network: Deploy from a stable connection. Retries use exponential backoff and should succeed automatically
  • If stuck in pending state:
    bun scripts/pages.ts list-deployments my-app  # Check status
    # If truly stuck, delete and recreate the project

Reference: Troubleshooting Guide

For detailed solutions to specific error messages, see the Troubleshooting section below.

Create and Use KV Storage

To create a KV namespace and store data:

# Create namespace
bun scripts/kv-storage.ts create-namespace user-sessions
# Returns: Namespace ID (e.g., abc123def456)
# Save this ID for binding to workers

# Write key-value pair
bun scripts/kv-storage.ts write <namespace-id> "session:user123" '{"userId":"123","token":"abc"}'

# Read value
bun scripts/kv-storage.ts read <namespace-id> "session:user123"
# Returns: {"userId":"123","token":"abc"}

# List all keys (useful for debugging)
bun scripts/kv-storage.ts list-keys <namespace-id>

# Delete a key
bun scripts/kv-storage.ts delete <namespace-id> "session:user123"

Important: KV storage uses eventual consistency. Writes may take up to 60 seconds to propagate globally. For immediate reads, use the same edge location where you wrote the data.

Create R2 Bucket and Upload Files

To create an R2 bucket and manage objects:

# Create bucket
bun scripts/r2-storage.ts create-bucket media-assets

# Upload file
bun scripts/r2-storage.ts upload media-assets ./images/logo.png logo.png

# List objects
bun scripts/r2-storage.ts list-objects media-assets

# Download object
bun scripts/r2-storage.ts download media-assets logo.png ./downloaded-logo.png

Deploy to Cloudflare Pages

To deploy a static site or application to Pages:

# Create Pages project (or get existing project info)
bun scripts/pages.ts deploy my-app ./dist
# Returns: https://my-app.pages.dev

# Set environment variable
bun scripts/pages.ts set-env my-app API_URL https://api.example.com

# Set environment variable for specific environment
bun scripts/pages.ts set-env my-app DEBUG true --env preview

# Get deployment URL
bun scripts/pages.ts get-url my-app

Auto-extracted URLs: The Pages script automatically extracts and returns the Cloudflare-generated URL (e.g., https://my-app.pages.dev) from the deployment response.

Note: The API creates the project structure, but for actual file uploads, you'll need Wrangler CLI:

npx wrangler pages deploy ./dist --project-name=my-app

Why this works: The skill creates/verifies the Pages project and returns the URL. For the initial deployment with files, Wrangler handles the complex multipart upload process.

Configure DNS and Routes

To create DNS records and configure worker routes:

# Create DNS A record
bun scripts/dns-routes.ts create-dns example.com A api 192.168.1.1

# Route pattern to worker
bun scripts/dns-routes.ts create-route example.com "*.example.com/api/*" api-handler

Common Workflows

Multi-Service Setup

To set up a complete application with worker, KV storage, and R2 bucket:

  1. Create KV namespace for caching
    bun scripts/kv-storage.ts create-namespace app-cache
  2. Create R2 bucket for media
    bun scripts/r2-storage.ts create-bucket app-media
  3. Deploy worker with bindings
    bun scripts/workers.ts deploy app-worker ./worker.js --kv-binding app-cache --r2-binding app-media
  4. Configure route
    bun scripts/dns-routes.ts create-route example.com "example.com/*" app-worker

Update Worker Configuration

To update an existing worker's code or bindings:

# Update worker code
bun scripts/workers.ts update worker-name ./new-worker-script.js

# Get worker details
bun scripts/workers.ts get worker-name

# List all workers
bun scripts/workers.ts list

Bulk KV Operations

To perform bulk operations on KV storage:

# Bulk write from JSON file
bun scripts/kv-storage.ts bulk-write namespace-name ./data.json

# Delete multiple keys
bun scripts/kv-storage.ts bulk-delete namespace-name key1 key2 key3

Error Handling

Missing API Key

If .env file is missing or CLOUDFLARE_API_KEY is not set:

Error: CLOUDFLARE_API_KEY not found in environment

Solution: Create .env file in project root:
  echo "CLOUDFLARE_API_KEY=your_token_here" > .env

Invalid Permissions

If API token lacks required permissions:

Error: Insufficient permissions for Workers deployment

Required: Workers Scripts: Edit
Current: Workers Scripts: Read

Solution: Update token permissions at:
  https://dash.cloudflare.com/profile/api-tokens

API Rate Limiting

If too many requests are made:

Error: Rate limit exceeded (429)

Solution: Retry automatically with exponential backoff (3 attempts)

Network Issues

If API is unreachable:

Error: Failed to connect to Cloudflare API

Solution: Check internet connection and retry

Best Practices

Security:

  • Never commit .env files - always add to .gitignore
  • Use token-based authentication (not API keys)
  • Rotate tokens periodically (every 90 days recommended)
  • Use least-privilege principle: only grant required permissions
  • Store secrets via Wrangler CLI: wrangler secret put SECRET_NAME

Performance:

  • Deploy workers to minimize latency (they run at Cloudflare edge)
  • Use KV storage for frequently-read data (not frequently-written)
  • Use R2 for large files (KV has 25MB limit per key)
  • Enable caching with appropriate TTLs
  • Keep worker scripts under 1MB for faster cold starts

Development Workflow:

  • Test locally first: wrangler dev for local testing
  • Use staging environment before production
  • Validate credentials after token updates: bun scripts/validate-api-key.ts
  • Monitor worker logs: wrangler tail worker-name
  • Version your workers: use names like api-v1, api-v2

Naming Conventions:

  • Workers: Use descriptive names (e.g., user-auth-worker not worker1)
  • KV namespaces: Include purpose (e.g., app-sessions, api-cache)
  • R2 buckets: Use lowercase with hyphens (e.g., media-assets-prod)
  • Be consistent across your infrastructure

Resource Management:

  • Delete unused workers, namespaces, and buckets
  • Monitor usage in Cloudflare dashboard
  • Free tier limits: 100,000 requests/day for Workers
  • Set up billing alerts to avoid surprises

Advanced Usage

For advanced scenarios including:

  • Complex routing configurations
  • Multi-region deployments
  • Custom domain setup
  • Worker-to-worker communication
  • Durable Objects integration
  • Bulk operations and migrations

See examples.md for comprehensive examples and patterns.

Script Reference

All scripts are located in ~/.claude/skills/cloudflare-manager/scripts/:

  • validate-api-key.ts: Validate API credentials and display permissions
  • workers.ts: Deploy, update, and manage Workers
  • kv-storage.ts: Create and manage KV namespaces and key-value pairs
  • r2-storage.ts: Create and manage R2 buckets and objects
  • pages.ts: Deploy and configure Cloudflare Pages projects
  • dns-routes.ts: Configure DNS records and worker routes
  • utils.ts: Shared utilities for API calls and error handling

Templates

Starter templates are available in ~/.claude/skills/cloudflare-manager/templates/:

  • worker-template.js: Basic worker template with fetch handler
  • wrangler.toml.template: Wrangler configuration template

Troubleshooting

Quick Link: High-Priority Issues

For the most critical problems encountered in production, see Common Pitfalls above:

  • API token scoping issues → API Token Scoping
  • Wrangler OAuth failures → Wrangler OAuth Port Conflicts
  • Deployment timeouts → Deployment Timeouts

Diagnostic Commands

First steps when something fails:

# 1. Validate API credentials and permissions
bun scripts/validate-api-key.ts

# 2. List existing resources to check state
bun scripts/workers.ts list
bun scripts/kv-storage.ts list-namespaces
bun scripts/r2-storage.ts list-buckets
bun scripts/pages.ts list-projects

# 3. Check specific resource details
bun scripts/workers.ts get worker-name
bun scripts/pages.ts list-deployments project-name

When to use each:

  • Step 1 (validate): Always run this first if ANY command fails
  • Step 2 (list): Check what resources actually exist
  • Step 3 (get details): Investigate specific resource state

Common Issues and Solutions

Issue: "Worker deployment failed with unknown error"

Symptoms: Deployment command exits with error code 1, no specific error message

Solutions:

  1. Check script syntax: node --check ./worker.js
  2. Verify file exists: ls -lh ./worker.js
  3. Re-validate API key: bun scripts/validate-api-key.ts --no-cache
  4. Check worker name is valid (alphanumeric, hyphens, underscores only)

Issue: "KV namespace not found"

Symptoms: Error when trying to read/write to namespace

Solutions:

  1. List all namespaces: bun scripts/kv-storage.ts list-namespaces
  2. Verify you're using namespace ID (not name) in commands
  3. Check namespace wasn't deleted
  4. Ensure API token has KV Storage permissions

Issue: "R2 bucket already exists" or "Bucket name taken"

Symptoms: Cannot create bucket with chosen name

Solutions:

  1. Bucket names must be globally unique across all Cloudflare accounts
  2. Try a more specific name: my-app-media-2024 instead of media
  3. Use existing bucket: bun scripts/r2-storage.ts list-buckets
  4. Names must be 3-63 characters, lowercase letters/numbers/hyphens only

Issue: "Pages deployment timeout" or "Deployment pending"

Symptoms: Deployment doesn't complete, stays in pending state

Solutions:

  1. Check deployment status: bun scripts/pages.ts list-deployments project-name
  2. View in dashboard: https://dash.cloudflare.com/pages
  3. Large deployments (>1000 files) may take 5-10 minutes
  4. Cancel and retry if stuck: Delete project and recreate

Issue: "DNS record creation failed"

Symptoms: Cannot create DNS records or routes

Solutions:

  1. Verify zone exists: bun scripts/dns-routes.ts list-zones
  2. Ensure domain is added to Cloudflare and active
  3. Check nameservers point to Cloudflare: dig NS yourdomain.com
  4. Verify API token has Zone > DNS > Edit permission

Issue: "API rate limit exceeded (429)"

Symptoms: Commands fail with "Too many requests"

Solutions:

  1. Scripts automatically retry with exponential backoff
  2. Wait 1-2 minutes before retrying manually
  3. Reduce concurrent operations
  4. Rate limits: 1200 requests per 5 minutes

Issue: "CLOUDFLARE_API_KEY not found in environment"

Symptoms: Commands fail immediately with environment error

Solutions:

  1. Create .env file in project root (not skill directory)
  2. Verify file content: cat .env | grep CLOUDFLARE_API_KEY
  3. Ensure no extra spaces: CLOUDFLARE_API_KEY=token (no spaces around =)
  4. Run commands from project root where .env exists

Quick Fix:

cd /path/to/your/project
echo "CLOUDFLARE_API_KEY=your_token_here" > .env
bun scripts/validate-api-key.ts

Security Notes

  • API keys are never logged or displayed in output
  • All API requests use HTTPS
  • User inputs are validated before API calls
  • Destructive operations (delete) require confirmation
  • Permissions are cached for 24 hours to minimize token exposure

Additional Resources

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

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.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.