Cloudflare MCP Server & CLI

What is the Cloudflare MCP server & CLI?

The Cloudflare MCP server & CLI is a free, open source program that lets AI agents read zone/DNS information and perform specifically approved management requests for you, in 2 ways. The MCP server is what an AI app like Claude, Codex or Cursor connects to, through MCP (Model Context Protocol), the open standard AI apps use to call outside tools.

You ask in plain language. Your AI picks the right tool, and the server makes the call directly to the fixed Cloudflare API origin.

The CLI is the same program as commands. cloudflare-cli list-zones runs the same code your AI runs when you ask to inspect the zones accessible to your private token, whether an agent like Claude Code runs it or you do.

What can you ask it?

Once it's set up, you ask the way you'd ask an assistant. These are real prompts it handles:

Try asking
Read DNS for this exact zone without changing records.
Show the complete current schema for a DNS record.
Preview this small DNS batch locally.
Apply only the same reviewed zone, profile and JSON body.
Read the selected zone setting before approving a change.
Inspect current Rulesets and their entrypoint.
Query the eligible analytics dataset for this time window.

Cloudflare already offers Code Mode and service MCPs, the general cf CLI and Wrangler. This owned package adds consistent local mandatory confirmation, isolated private profiles, no automatic request replay and an exact-request DNS batch review. Official coverage is much broader; current pinned comparisons acknowledge profiles, dry-run, delete confirmation and sandboxed credential handling.

How to install the Cloudflare MCP server

Use the existing install box for local MCP, the shared CLI or the versioned desktop bundle. Codex is the primary documented agent; complete client and OS setup follows below.

Set up Cloudflare access

Private API tokens and resource access

1. Choose the intended Cloudflare user/account and zone. Create a least-privilege API token through My Profile > API Tokens, or Manage account > Account API tokens for an account-owned token.
2. Choose only the permissions needed below and restrict account/zone resources. Check optional expiry and IP filters. Save the secret privately outside every repository; never use the Global API Key with this Bearer-only package.
3. Configure CLOUDFLARE_TOKEN_FILE as an absolute token-only file, or privately set CLOUDFLARE_API_TOKEN. Optional CLOUDFLARE_ACCOUNT_ID and CLOUDFLARE_ZONE_ID supply input defaults. A default does not restrict token permissions.
4. Run cloudflare-cli doctor for local configuration checks. Deliberately run doctor --network to verify token status: user tokens use GET /user/tokens/verify; account-owned tokens need CLOUDFLARE_TOKEN_KIND=account and a private account ID, using GET /accounts/{account_id}/tokens/verify. Only an active result is accepted; token IDs/status bodies are not printed.
5. Read the exact zone and DNS records, inspect the complete current input schema, preview the intended change locally and approve only the chosen request. A local preview does not verify provider permissions, conflicts or current remote state.

Requests use Authorization: Bearer on the fixed https://api.cloudflare.com/client/v4 origin. Token-kind selection changes the diagnostic route, not ordinary authorization or resource access. The wrapper does not implement OAuth login, refresh, saved official cf sessions, automatic .env loading or global CLI config ingestion. login prints instructions; it neither opens consent nor saves credentials.

User tokens inherit permitted user access. Account-owned tokens act as service principals; creation requires the provider's token-provisioning capability or Super Administrator role and grants cannot exceed the creator's permissions. Some products still have compatibility restrictions; verify the exact operation in the account-token compatibility documentation.

Review API token permissions and the selected endpoint's accepted grant list; this table is task guidance, not a complete static authorization manifest. Do not grant token-management, account administration or every zone solely for installation.

Use a private 0700 directory and regular 0600 token-only file on macOS/Linux. Windows users must restrict the file ACL to their own user; POSIX checks do not establish Windows ACL protection. Token files cannot be symlinks or exceed 64 KiB. File credentials override environment credentials and are cached until restart. GUI applications may have a different environment from your terminal.

Plans and limits

This AGPL wrapper is free; Cloudflare services and plan restrictions remain separate. DNS, proxy, cache purge, setting values, Ruleset phases and analytics availability are controlled by Cloudflare. Read metadata for the selected setting or dataset rather than assuming every write is allowed on every plan.

Current general API limits list 1200 requests per five minutes for a user/account token and 200 per second per IP, with cumulative user limits and endpoint-specific exceptions. Other sessions and processes consume the same upstream quota. Local pacing defaults to 200 ms per profile/process; it does not reserve quota. No reads or mutations retry automatically after 429, timeout, network failure or HTTP 200 provider errors. Wait according to current provider response/reset policy and inspect state before a deliberate repeat.

GraphQL analytics limits separately describe a default 300-query/five-minute user quota, up to ten zones or one account per scoped query, and dataset-specific retention/record limits. The general limits table uses a different GraphQL maximum figure; use the current analytics rules and returned policy for the actual query. Account-based limiting can be enabled through provider controls; the wrapper does not opt you into it. Inspect the provider's settings node for your dataset and resource. Adaptive analytics may be sampled; an aggregate is not necessarily an exact event ledger.

Reviewed DNS batches have a local cap of 1–50 total actions, deliberately below many provider plan limits. The underlying native batch command uses the reviewed source schema and provider limits. Request JSON is capped at 1 MiB; responses at 5 MiB; GraphQL query text at 64 KiB and 10000 parsed tokens; default timeout is 30 seconds. These caps do not raise Cloudflare quotas or plan allowances.

Rotation and revocation

Revoke or rotate the selected user/account token in its Cloudflare settings, update private configuration and restart every client using it. npm uninstall does not revoke the token, restore DNS, remove Rulesets or undo a cache purge. Keep account exports, DNS TXT values, signed URLs and diagnostic responses out of public issues.

Check that it works

Start with local configuration, then deliberately verify the configured user/account token route and active status without returning token details.

cloudflare-cli --version
cloudflare-cli doctor
cloudflare-cli doctor --network
cloudflare-cli list-accounts --agent
cloudflare-cli list-zones --per-page 10 --agent

cloudflare-cli doctor
cloudflare-cli doctor --network
cloudflare-cli list-accounts --agent
cloudflare-cli list-zones --per-page 10 --agent
cloudflare-cli get-zone --zone-id YOUR_ZONE_ID --agent

Local doctor reports configuration, not credentials validated. --network checks an active token through its configured token-kind route; it does not prove access to every endpoint. First provider read should be the exact zone needed for the task. Do not create DNS records or purge cache merely to verify installation.

Full discovery has 29 tools; CLOUDFLARE_READ_ONLY=1 hides eleven mutations and direct calls still refuse. Local helpers expose no tokens or provider IDs. A missing credential exits 10; a valid install is not proof of provider account access.

Use the Cloudflare CLI

The CLI is the same 29 tools as commands. AI agents that run commands, like Claude Code, Codex and OpenCode, use it on their own, and you can type the same commands in a terminal or a script.

Every tool name becomes a command with dashes, so get_dns_record runs as cloudflare-cli get-dns-record.

cloudflare-cli tools
cloudflare-cli create-dns-record --help
cloudflare-cli schema create-dns-record
cloudflare-cli list-zones --per-page 10 --agent
cloudflare-cli get-zone --zone-id YOUR_ZONE_ID --agent

The bare cloudflare-cli lists every command, and cloudflare-cli <command> --help shows what a command takes. All eleven mutations require --confirm. --agent/--yes do not approve changes. Use a complete native payload or payload_file; local previews do not validate provider permissions or state.

These flags work on every command:

A script can branch on the exit code:

MCP server or CLI: which one?

Both surfaces call the same tools. Codex can connect to the local MCP server or run the CLI directly. Neither requires Claude Code.

MCP provides structured tool discovery; the CLI supports scripts, compact JSON, field selection and command/schema discovery. Official hosted MCP connection and local CLI authentication have different setup requirements.

Codex-specific token measurements are pending. Record the actual client/model versions, discovery configuration, input/output usage, caching, latency and equivalent successful outcomes. Standing definitions and full task cost are separate measurements; CLI commands, selected help, results and reasoning still consume tokens.

No efficiency percentage or Claude-derived figure is presented as a Codex result. Other-client benchmarks can be added separately.

DNS, Rulesets, cache and analytics workflows

A deliberate DNS change

Read list_zones, get_zone and list_dns_records to establish the exact zone and record. Inspect get_operation_schema for create_dns_record or update_dns_record. A PATCH edits provided fields; a PUT replacement can alter omitted-field behavior. Choose ttl/proxied/type/content deliberately rather than assuming every record should be proxied. DNS conflicts, plan constraints and semantic validity are finally decided by Cloudflare.

cloudflare-cli preview-operation --operation create_dns_record --arguments '{"zone_id":"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa","payload":{"name":"review.example.com","type":"A","content":"192.0.2.1","ttl":1,"proxied":false}}' --agent
cloudflare-cli create-dns-record --zone-id YOUR_ZONE_ID --payload-file /absolute/private/approved-dns.json --account work --confirm --agent

The first command is a local illustration using documentation-only IDs/IPs. It makes no provider request. The second is a real mutation only when private credentials, a real resource and the explicitly approved body are supplied. After any approved change, read the returned record and check the intended resolver/service.

Review a small DNS batch

Save only the chosen native deletes/patches/puts/posts body in a private JSON file. preview_dns_batch validates the same native schema and 1–50 total-action cap, then returns a SHA-256 digest of method, exact zone path, profile label and canonically sorted JSON object keys. Array order is preserved. Read and approve the full body, target and profile before apply_dns_batch.

cloudflare-cli preview-dns-batch --zone-id YOUR_ZONE_ID --payload-file /absolute/private/approved-dns-batch.json --account work --agent
cloudflare-cli apply-dns-batch --zone-id YOUR_ZONE_ID --payload-file /absolute/private/approved-dns-batch.json --account work --preview-sha256 REVIEWED_64_CHARACTER_HASH --confirm --agent

Any change to that reviewed request requires a new preview. This does not detect concurrent remote edits, credential rotation under the same label or changes in token grants. The digest is not an authorization secret, human approval signature or remote-state transaction lock. The direct native batch_dns_records command also exists, requires --confirm and does not require a digest; choose the review/apply workflow when you need exact local review.

Cloudflare batch semantics execute deletes, patches, puts, then posts in a database transaction; distributed DNS propagation remains non-atomic. If provider validation fails, no batch changes apply. A network failure after submission can leave the outcome unknown; read the affected records before repeating. The wrapper never promises all resolvers change at once or supplies a rollback.

Settings and Rulesets

get_zone_setting with setting_id=ssl replaces the older SSL helper. Inspect the setting's editable/value metadata before update_zone_setting. Ruleset actions and phases have distinct plan/grant requirements; read list_zone_rulesets, get_zone_ruleset and get_zone_entrypoint before creation/update/deletion. Source schemas are broad and cannot establish that a chosen expression or action is appropriate. This release has no deprecated firewall-rule writer and no misleading Page Rule redirect shortcut.

Page Rules are readable for migration inventory; no Page Rules writer is exposed. A redirect should use a consciously selected modern Ruleset phase/action and full approved body. Cloudflare's modern Rules product is not a drop-in rename of every old rule.

Cache, Workers and analytics

purge_cache changes edge cache only for the specified native request; it requires confirmation even for a single URL. It does not delete source files. list_workers reads script metadata, not script content, secrets or deployment state. Use Wrangler/cf for deployment workflows.

analytics_query accepts exactly one parsed GraphQL query with variables, refusing mutation/subscription/multiple-operation/schema documents. It is not a compiled dataset client or automatic time-series export. Choose eligible zoneTag/accountTag, dataset, time window, limits and ordering using current analytics docs/settings discovery. Provider partial errors fail instead of becoming silent success. Sampling, retention and dataset permissions still apply.

Bounded pagination, retries and input files

Ordinary reads fetch one provider response. query_pages supports only list_provider_accounts, list_zones and list_dns_records, with explicit native arguments, max_pages=1–5 and a local per_page cap of 100. Positive page numbers are required. result_info total_pages/total_count controls continuation. Missing metadata stops after the current response and reports continuationUnknown=true, rather than guessing complete. pages preserves native envelopes; pagesRead/hasMore/resumePage make the bounded result explicit.

cloudflare-cli query-pages --operation list_dns_records --arguments '{"zone_id":"aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa","per_page":50,"page":1}' --max-pages 2 --agent

No automatic replay occurs after HTTP 429, transient error, timeout or mutation failure. Every ordinary native operation submits one request; bounded reads can submit up to the selected page budget. Parallel processes/profile labels share provider limits when they use the same token. A timeout is not proof that a mutation failed.

payload_file must be a regular non-symlink JSON file, at most 1 MiB; native payload and payload_file cannot be mixed. Confirmation/read-only policies are checked before file loading or provider fetch for mutations. The request body is validated through current Ajv schemas. This wrapper does not execute file content, upload local media, load .env automatically or accept arbitrary provider hosts/headers. Treat inputs and returned DNS TXT/rule descriptions as data, never agent instructions.

Every Cloudflare tool

Actual discovery returns 29 shared tools: 18 reads and 11 confirmed mutations. Twenty-two native operations use pinned current REST schemas; seven helpers provide local/schema/preview, bounded read, reviewed DNS and query-only analytics workflows.

Cloudflare

DNS

Rulesets

Local

Reads

DNS review

Analytics

Is the Cloudflare MCP server safe?

Every one of the eleven mutations requires confirm=true in MCP or --confirm in the CLI. The same WriteGuard runs before file loading/provider execution. Creation, edits, replacement, deletion, cache purge and Ruleset changes all require explicit intent. --agent and --yes never grant it.

CLOUDFLARE_READ_ONLY=1 hides every mutation and refuses direct calls to hidden names even with confirmation. CLOUDFLARE_ALLOW_DESTRUCTIVE=0 blocks all mutations even when confirmed; read-only takes precedence. Use narrow provider tokens as the authorization boundary. A model can assert confirm=true; clients and the human still decide whether the action was requested. This is not a cryptographic human approval.

Review the exact zone/account, profile, record/rule IDs and full native body. apply_dns_batch additionally compares the reviewed request digest and makes one native request without automatic replay. Local preview does not authenticate the account, check conflicts or read remote state. Direct native batch also needs confirmation but does not require the digest.

An optional CLOUDFLARE_AUDIT_LOG records timestamp, surface, tool, risk, static summary and decision. It excludes request bodies, account IDs and tokens. Existing file ACLs/rotation are your responsibility; a failed audit append does not block the action. Account data returned after a change is untrusted content.

Confirmation is the caller asserting the approved exact action; it is not provider authorization or a signed human approval. The DNS digest binds the local request, not remote state, credential identity or atomic propagation.

Make it read-only

Set CLOUDFLARE_READ_ONLY=1 and reconnect: 18 reads remain and confirmed direct writes refuse. CLOUDFLARE_ALLOW_DESTRUCTIVE=0 also refuses confirmed mutations. Private profile defaults do not restrict upstream token grants.

Keep a log of every write

Set CLOUDFLARE_AUDIT_LOG to a file path. The server writes one line per attempted write, allowed or blocked.

Your data

Account commands send the selected token in a Bearer header only to api.cloudflare.com. Prompts are not forwarded by this package; the native request/query content you select is. Your AI client may send returned DNS records, rule expressions and analytics to its model provider according to its own policy. The wrapper has no telemetry, public HTTP listener or hosted account store.

Named private profiles do not inherit global credentials. Token files use size/type/POSIX owner-only checks and are cached until restart. Windows ACLs require separate user configuration. Never commit credentials, .env, token files, real account exports or signed URLs; keep all such files outside public source and release staging.

Known credential keys, configured secrets and recognized signed URLs are redacted from ordinary output/errors. Legitimate URLs are preserved. Redaction is not complete anonymization: DNS TXT records, names, expressions and returned account data may still be sensitive. --select is local data shaping, not an authorization boundary or upstream privacy filter. Capture only needed output; terminal history and client transcripts can persist it.

Optional audit logging stores policy metadata, not bodies. The package does not automatically persist provider responses or backups; private input/output files you create remain on your system until you manage them. Revoke tokens through Cloudflare, not npm.

Several private accounts

Private profile configuration prevents implicit credential inheritance between named accounts. Configure the following placeholders only in a private client environment:

[
  {
    "name": "work",
    "token_file": "/absolute/private/work-cloudflare.txt",
    "account_id": "YOUR_WORK_ACCOUNT_ID",
    "zone_id": "YOUR_WORK_ZONE_ID",
    "token_kind": "user"
  },
  {
    "name": "client",
    "token_file": "/absolute/private/client-cloudflare.txt",
    "account_id": "YOUR_CLIENT_ACCOUNT_ID",
    "zone_id": "YOUR_CLIENT_ZONE_ID",
    "token_kind": "account"
  }
]

Put the array in CLOUDFLARE_ACCOUNTS and set CLOUDFLARE_DEFAULT_ACCOUNT=work. Names must be unique and nonempty. --account selects the exact label; list_accounts returns only label/default/auth method, never token values, file paths or provider IDs. A profile lacking credentials does not fall back to a global token. Global account/zone defaults likewise do not leak into explicitly named entries.

account_id/zone_id are routing defaults; explicit operation inputs take precedence and can name another resource allowed by the token. Token-kind defaults to user and only changes the diagnostic verification route. Use provider token resource scopes for real access boundaries. Local profile separation is not a multi-user access-control service.

Cloudflare MCP server settings

Configure these in private shell/user-client settings; no automatic .env loading or global official CLI credentials are used.

Troubleshooting

Run the doctor first. It names the step that failed and the fix.

If the server doesn't show up in your app at all, run the command your app runs, in a terminal, and read the error.

Every argument and nested native input

Twenty-two current native operations and seven helpers. Each schema below comes from actual stdio discovery. Native payload bodies preserve required keys, enums, unions, nested references and source constraints. Missing zone_id/account_id may use the selected private profile default; other required arguments remain required. Confirmation is enforced outside ordinary schema-required lists.

list_provider_accounts

cloudflare-cli list-provider-accounts

List all accounts you have ownership or verified access to. Read operation.

Native operation: GET /accounts; accounts-list-accounts. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

list_zones

cloudflare-cli list-zones

Lists, searches, sorts, and filters your zones. Listing zones across more than 500 accounts is currently not allowed. Read operation.

Native operation: GET /zones; zones-get. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

get_zone

cloudflare-cli get-zone

Retrieves detailed information about a specific zone identified by its zone ID.

Returns zone configuration, status, nameservers, and associated metadata. Read operation.

Native operation: GET /zones/{zone_id}; zones-0-get. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

list_dns_records

cloudflare-cli list-dns-records

List, search, sort, and filter a zones' DNS records. Read operation.

Native operation: GET /zones/{zone_id}/dns_records; dns-records-for-a-zone-list-dns-records. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

get_dns_record

cloudflare-cli get-dns-record

Retrieves details for a specific DNS record in the zone. Read operation.

Native operation: GET /zones/{zone_id}/dns_records/{dns_record_id}; dns-records-for-a-zone-dns-record-details. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

create_dns_record

cloudflare-cli create-dns-record

Create a new DNS record for a zone.

Notes: - A/AAAA records cannot exist on the same name as CNAME records. - NS records cannot exist on the same name as any other record type. - Domain names are always represented in Punycode, even if Unicode characters were used when creating the record. Every change requires explicit confirmation; never repeat an unknown outcome automatically.

Native operation: POST /zones/{zone_id}/dns_records; dns-records-for-a-zone-create-dns-record. Native body required through payload or payload_file. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

update_dns_record

cloudflare-cli update-dns-record

Update an existing DNS record.

Notes: - A/AAAA records cannot exist on the same name as CNAME records. - NS records cannot exist on the same name as any other record type. - Domain names are always represented in Punycode, even if Unicode characters were used when creating the record. Every change requires explicit confirmation; never repeat an unknown outcome automatically.

Native operation: PATCH /zones/{zone_id}/dns_records/{dns_record_id}; dns-records-for-a-zone-patch-dns-record. Native body required through payload or payload_file. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

overwrite_dns_record

cloudflare-cli overwrite-dns-record

Overwrite an existing DNS record.

Notes: - A/AAAA records cannot exist on the same name as CNAME records. - NS records cannot exist on the same name as any other record type. - Domain names are always represented in Punycode, even if Unicode characters were used when creating the record. Every change requires explicit confirmation; never repeat an unknown outcome automatically.

Native operation: PUT /zones/{zone_id}/dns_records/{dns_record_id}; dns-records-for-a-zone-update-dns-record. Native body required through payload or payload_file. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

delete_dns_record

cloudflare-cli delete-dns-record

Permanently removes a DNS record from the zone. Every change requires explicit confirmation; never repeat an unknown outcome automatically.

Native operation: DELETE /zones/{zone_id}/dns_records/{dns_record_id}; dns-records-for-a-zone-delete-dns-record. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

batch_dns_records

cloudflare-cli batch-dns-records

Send a Batch of DNS Record API calls to be executed together.

Notes: - Although Cloudflare will execute the batched operations in a single database transaction, Cloudflare's distributed KV store must treat each record change as a single key-value pair. This means that the propagation of changes is not atomic. See [the documentation](https://developers.cloudflare.com/dns/manage-dns-records/how-to/batch-record-changes/ "Batch DNS records") for more information. - The operations you specify within the /batch request body are always executed in the following order:

- Deletes - Patches - Puts - Posts Every change requires explicit confirmation; never repeat an unknown outcome automatically.

Native operation: POST /zones/{zone_id}/dns_records/batch; dns-records-for-a-zone-batch-dns-records. Native body required through payload or payload_file. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

purge_cache

cloudflare-cli purge-cache

Deletes cached content in every Cloudflare data center and cache tier, including Cache Reserve. The next request for purged content is a cache MISS: Cloudflare fetches the full response from your origin and caches it again. Cloudflare does not serve purged content from cache again, even if your origin is unavailable.

To keep content cached and have Cloudflare revalidate it with your origin instead, use POST /zones/{zone_id}/invalidate_cache.

Choose what to purge

Send one of these fields in the request body:

• files: specific URLs. If your cache key includes request headers, send each URL with the header values it was cached with.
• tags: all content whose Cache-Tag response header contains one of the tags.
• hosts: all content cached for the hostnames.
• prefixes: all content whose URL starts with one of the prefixes.
• purge_everything: all cached content in the zone.

Check the result

A 200 response with success: true means Cloudflare accepted the request. It does not confirm that any content was cached or removed. To check, request a purged URL and confirm that the CF-Cache-Status response header is MISS.

Availability and limits

Rate limits and the number of items you can send in one request depend on your plan. See Purge cache: availability and limits. Every change requires explicit confirmation; never repeat an unknown outcome automatically.

Native operation: POST /zones/{zone_id}/purge_cache; zone-purge. Native body required through payload or payload_file. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

get_zone_setting

cloudflare-cli get-zone-setting

Fetch a single zone setting by name Read operation.

Native operation: GET /zones/{zone_id}/settings/{setting_id}; zone-settings-get-single-setting. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

update_zone_setting

cloudflare-cli update-zone-setting

Updates a single zone setting by the identifier Every change requires explicit confirmation; never repeat an unknown outcome automatically.

Native operation: PATCH /zones/{zone_id}/settings/{setting_id}; zone-settings-edit-single-setting. Native body required through payload or payload_file. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

list_workers

cloudflare-cli list-workers

Fetch a list of uploaded Worker scripts. Read operation.

Native operation: GET /accounts/{account_id}/workers/scripts; worker-script-list-workers. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

list_zone_rulesets

cloudflare-cli list-zone-rulesets

Fetches all rulesets at the zone level. Read operation.

Native operation: GET /zones/{zone_id}/rulesets; listZoneRulesets. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

get_zone_ruleset

cloudflare-cli get-zone-ruleset

Fetches the latest version of a zone ruleset. Read operation.

Native operation: GET /zones/{zone_id}/rulesets/{ruleset_id}; getZoneRuleset. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

get_zone_entrypoint

cloudflare-cli get-zone-entrypoint

Fetches the latest version of the zone entry point ruleset for a given phase. Read operation.

Native operation: GET /zones/{zone_id}/rulesets/phases/{ruleset_phase}/entrypoint; getZoneEntrypointRuleset. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

create_zone_ruleset

cloudflare-cli create-zone-ruleset

Creates a ruleset at the zone level. Every change requires explicit confirmation; never repeat an unknown outcome automatically.

Native operation: POST /zones/{zone_id}/rulesets; createZoneRuleset. Native body required through payload or payload_file. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

update_zone_ruleset

cloudflare-cli update-zone-ruleset

Updates a zone ruleset, creating a new version. Every change requires explicit confirmation; never repeat an unknown outcome automatically.

Native operation: PUT /zones/{zone_id}/rulesets/{ruleset_id}; updateZoneRuleset. Native body required through payload or payload_file. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

delete_zone_ruleset

cloudflare-cli delete-zone-ruleset

Deletes all versions of an existing zone ruleset. Every change requires explicit confirmation; never repeat an unknown outcome automatically.

Native operation: DELETE /zones/{zone_id}/rulesets/{ruleset_id}; deleteZoneRuleset. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

list_page_rules

cloudflare-cli list-page-rules

Fetches Page Rules in a zone. Read operation.

Native operation: GET /zones/{zone_id}/pagerules; page-rules-list-page-rules. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

get_page_rule

cloudflare-cli get-page-rule

Fetches the details of a Page Rule. Read operation.

Native operation: GET /zones/{zone_id}/pagerules/{pagerule_id}; page-rules-get-a-page-rule. Native body not required. Schema discovery is local; endpoint permissions/plan decisions remain provider controlled.

list_accounts

cloudflare-cli list-accounts

Local profile labels/default/auth method only. No provider IDs, credential values or paths. No network.

get_operation_schema

cloudflare-cli get-operation-schema

Complete selected current API input, path and query schema. Local metadata only.

preview_operation

cloudflare-cli preview-operation

Validate a named operation and return its exact local method/path/query/body/profile. No request, auth validation or provider policy checks.

query_pages

cloudflare-cli query-pages

Read at most five pages of a supported paginated named operation. Provider page metadata determines continuation; missing metadata stops with an explicit unknown continuation.

preview_dns_batch

cloudflare-cli preview-dns-batch

Local schema-validated DNS batch review capped at 50 actions. Digest binds exact JSON body, zone path and profile label; no account or DNS state read. Not a provider authorization grant.

apply_dns_batch

cloudflare-cli apply-dns-batch

Verify the reviewed digest and submit one exact native DNS batch with mandatory confirmation. Does not assert remote-state consistency or atomic DNS propagation; no retry.

analytics_query

cloudflare-cli analytics-query

One parsed GraphQL query only, with variables. No mutations, subscriptions or multiple operations. Permissions, dataset retention, sampling and query limits remain provider controlled.

Nested native input definitions

Identical object shapes are listed once below. References point to the named definitions; union branches retain their individual requirements. Some provider JSON-schema conditions do not fit a flat table; use schema COMMAND or get_operation_schema for complete validation. Unknown native root body keys are refused locally; provider constraints still apply.

##### list_provider_accounts

##### list_zones