Cloudflare MCP Skill

Interact with Cloudflare services via MCP: Workers, KV, R2, D1, Hyperdrive, Observability, Builds, Audit, Container Sandbox.

File-based Pipeline (Pass Paths Only)

When integrating troubleshooting/changes into multi-step workflows, persist all evidence and artifacts to disk, passing only paths between agents/sub-agents.

Recommended directory structure (within project): runs/<workflow>/active/<run_id>/

Input: 01-input/goal.md (symptoms/objectives), 01-input/context.json (account/worker/resource/time_range, etc.)
Evidence: 02-analysis/observability.md, 02-analysis/audit.md, 02-analysis/screenshots/
Plan: 03-plans/change-plan.md (write operation plan; must write here and await confirmation first)
Output: 05-final/report.md (conclusion + evidence chain + tool call summary + next steps)
Logs: logs/events.jsonl (summary of each tool call)

Permission Tiers (Core Principles)

Tier	Purpose	Tool Scope	Risk Control
Diagnose	Read-only/query/troubleshoot	Observability, Builds, Browser, Audit	Default entry point, no write operations
Change	Create/modify/delete resources	Workers Bindings (KV/R2/D1)	Requires user confirmation, post-execution verification
Super Admin	Highest privileges	All + Container Sandbox	Only in isolated environments/test accounts

Tool Reference

Diagnose Tier (Read-only)

Observability

Tool	Purpose
`query_worker_observability`	Query logs/metrics (events, CPU, error rate)
`observability_keys`	Discover available fields
`observability_values`	Explore field values

Builds

Tool	Purpose
`workers_builds_list_builds`	List build history
`workers_builds_get_build`	Get build details
`workers_builds_get_build_logs`	Get build logs

Browser Rendering (Page Capture)

Tool	Purpose
`get_url_html_content`	Fetch page HTML
`get_url_markdown`	Convert to Markdown
`get_url_screenshot`	Take page screenshot

Audit Logs

Tool	Purpose
`auditlogs_by_account_id`	Pull change history by time range

Change Tier (Write Operations)

Account

Tool	Purpose
`accounts_list`	List accounts
`set_active_account`	Set active account

Builds (Settings)

Tool	Purpose
`workers_builds_set_active_worker`	⚠️ Set active worker (requires confirmation)

Tool	Purpose
`kv_namespaces_list`	List namespaces
`kv_namespace_get`	Get details
`kv_namespace_create`	Create (⚠️ requires confirmation)
`kv_namespace_update`	Update (⚠️ requires confirmation)
`kv_namespace_delete`	Delete (⚠️ requires confirmation)

Tool	Purpose
`r2_buckets_list`	List buckets
`r2_bucket_get`	Get details
`r2_bucket_create`	Create (⚠️ requires confirmation)
`r2_bucket_delete`	Delete (⚠️ requires confirmation)

Tool	Purpose
`d1_databases_list`	List databases
`d1_database_get`	Get details
`d1_database_query`	Execute SQL
`d1_database_create`	Create (⚠️ requires confirmation)
`d1_database_delete`	Delete (⚠️ requires confirmation)

Hyperdrive

Tool	Purpose
`hyperdrive_configs_list`	List configs
`hyperdrive_config_get`	Get details
`hyperdrive_config_create`	Create (⚠️ requires confirmation)
`hyperdrive_config_edit`	Edit (⚠️ requires confirmation)
`hyperdrive_config_delete`	Delete (⚠️ requires confirmation)

Workers

Tool	Purpose
`workers_list`	List workers
`workers_get_worker`	Get worker details
`workers_get_worker_code`	Get source code

Super Admin Tier (Container Sandbox)

Tool	Purpose
`container_initialize`	Initialize container (~10 min lifecycle)
`container_exec`	Execute command
`container_file_write`	Write file
`container_file_read`	Read file
`container_files_list`	List files
`container_file_delete`	Delete file

Container Notes: No persistent state, short lifespan, only for temporary tasks (running tests/reproducing issues/parsing data).

Security Rules (Must Follow)

Read Operations

Define scope first: account / worker / resource ID
No account? Run accounts_list first
Conclusions must have evidence chain: logs/screenshots/audit records

Write Operations (Three-step Flow)

1. Plan: Read current state first (list/get)
2. Confirm: Output precise change (name/ID/impact scope), await user confirmation
3. Execute: create/delete/update
4. Verify: audit logs + observability confirm no new errors

Prohibited Actions

❌ Execute create/delete/update without confirmation
❌ Delete production resources (unless user explicitly says "delete production xxx")
❌ Use Super Admin privileges in non-isolated environments
❌ Use container sandbox as persistent environment

Operation Workflows

Troubleshooting Flow (Typical)

1. Clarify symptoms → worker name/time range/error type
2. query_worker_observability to pull logs/metrics
3. If build-related → workers_builds_get_build_logs
4. If page-related → get_url_screenshot to reproduce
5. Trace changes → auditlogs_by_account_id
6. Summarize: root cause + evidence + fix recommendations

Resource Management Flow

1. accounts_list → set_active_account
2. List resources (kv_namespaces_list / r2_buckets_list / d1_databases_list)
3. Plan changes → present to user
4. Execute after confirmation
5. Verify: audit logs + observability shows no errors

Output Format

Language: English
Structure: Conclusion → Key data/evidence → Tool call summary → Next steps
Write operations: Must clearly list operations to be executed and impact scope

Example output:

✅ Investigation complete: worker `api-gateway` experienced 5xx spike between 18:00-18:30

Root cause: New code deployed threw TypeError when processing /v2/users
Evidence:
- Logs: 18:02 first occurrence of "Cannot read property 'id' of undefined"
- Audit: 18:00 user dev@example.com deployed new version
- Metrics: error_rate jumped from 0.1% to 12%

Recommendation: Roll back to previous version, or fix /v2/users handler

Error Handling

Situation	Action
Account not set	Run `accounts_list` → `set_active_account` first
Resource doesn't exist	Verify ID/name is correct, list available resources
Insufficient permissions	Explain required permissions, suggest checking API token scope
Observability query too long	Split into smaller time ranges, ask more specific questions

Scenario Examples

See scenarios.md for 20 real-world development scenarios.

mcp-cloudflare

Safety Notice

Copy this and send it to your AI assistant to learn