本文へ移動
cccskills
無料GitHub で公開

cloudflare-one-migrations

Plans migrations from Zscaler ZIA/ZPA, Palo Alto, legacy VPN, SWG, or SASE stacks to Cloudflare One. Use for migration assessments, policy mapping, rollout plans, and parity/gap analysis.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md12.1 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Cloudflare One Migrations

Retrieve current Cloudflare docs, Cloudflare API schemas, and source-vendor export docs before generating exact configuration.

Workflow

  1. Identify the source stack: Zscaler ZIA, Zscaler ZPA, Palo Alto NGFW/Prisma/GlobalProtect, legacy VPN/SWG/SD-WAN, or other.
  2. Request exports and logs before mapping. Prefer structured exports over screenshots or prose summaries.
  3. Build an inventory: identities, groups, apps, destinations, connectors/tunnels, DNS/URL/firewall/DLP/TLS policies, objects/lists, locations/sites, exceptions, hit counts, and compliance logging.
  4. Produce a mapping plan: source object, Cloudflare One target resource, confidence, prerequisites, unsupported/partial mappings, and manual decisions.
  5. Create dependencies first: identity/SCIM, connectors/on-ramps, routes/DNS, lists/objects, TLS bypasses, Access apps/policies, Gateway policies, DLP/CASB, logging.
  6. Stage safely: use a migration prefix, create disabled/audit-mode rules by default, pilot with small groups/sites, compare logs, then expand rollout.
  7. Account for every source rule. Each rule must map to a Cloudflare object or an explicit Not Migrated row with reason and security impact.

Exports To Ask For

  • ZIA: URL filtering, firewall filtering, SSL inspection, DLP, custom URL categories, IP groups, network services/service groups, users/groups/departments, locations, GRE tunnels, and static IPs.
  • ZPA: app segments, segment groups, server groups, app connectors/connector groups, access policies, IdP/group mapping, private DNS domains, ports, and protocols.
  • Palo Alto/Prisma: security/NAT/decryption rules, address/service objects and groups, URL categories, HIP profiles, GlobalProtect config, Prisma Access remote network/service connection config, zones, tags, logs, and hit counts.

Mapping Heuristics

Migration Assessment Prompts

  • Source coverage: which products are in scope, which exports are available, and whether screenshots/prose summaries are hiding missing object files.
  • Rule volume and hit data: counts by rule type, disabled/stale rules, no-hit rules, high-hit rules, and business-critical exceptions.
  • Object dependencies: address objects, service objects, groups, custom categories, network services, app IDs, zones, tags, connectors, and server groups.
  • Identity readiness: IdP, SCIM/group sync, group-name normalization, individual-user rules, local groups, service accounts, and contractor identities.
  • TLS/DLP readiness: source decryption rules, certificate-pinned bypasses, DLP engines/profiles, custom regex, exact-match data, and payload logging expectations.
  • Connectivity readiness: source tunnels/connectors, private DNS, Split Tunnels or bypass behavior, source IP preservation, egress IP allowlists, and site-to-site requirements.
  • Rollout readiness: pilot groups/sites, parallel-run period, rollback owner, source-stack decommission criteria, and monitoring/log comparison plan.

Source-Specific Traps

Zscaler ZIA / SWG

  • Custom URL categories often split into separate IP, domain, and URL lists. Count the generated lists, not just source categories.
  • ZIA locations with IPs are useful as source IP lists; they are not automatically Gateway DNS locations for DNS policy scoping.
  • GRE tunnel source IPs can inform policy conditions, but the transport migration is a separate WARP Connector or Cloudflare WAN workstream.
  • CAUTION/warn behavior has no exact Gateway equivalent. Treat it as an explicit customer decision, not a silent allow/block choice.
  • DLP engines and custom regex usually require manual Cloudflare DLP profile recreation. Placeholder policies must not be enabled as if DLP is complete.
  • Network application groups and unsupported protocols are partial mappings. Review them before enablement.
  • If SCIM is unavailable, identity-scoped source rules become overly broad unless you add an enforceable alternative such as user/email lists. Check Gateway identity selectors before creating those rules.

Zscaler ZPA / Private Access

  • ZPA app segments, server groups, and connector groups do not map 1:1. Cloudflare separates Access apps, tunnel routes, DNS, and policies.
  • Creating tunnels through the API does not complete connector deployment. Plan cloudflared installation, authentication, and origin reachability separately.
  • Create one Cloudflare Tunnel per ZPA connector group regardless of connector runtime status (AUTHENTICATED, DISCONNECTED, or disabled). Status is operational, not architectural. Tag disconnected or legacy groups in the tunnel description and let the customer decide what to decommission after validation.
  • Each ZPA connector instance within a group maps to one cloudflared replica running against that tunnel's token. Match replica count to connector instance count per group to preserve the same topology. A single tunnel token supports multiple simultaneous cloudflared processes. Recommend installing replicas within the same data center but on different hosts or subnets.
  • For each connector group, identify all server groups linked to it and all app segments assigned to those server groups. IP addresses and CIDRs in those app segments become CIDR routes on the corresponding tunnel; domain names become hostname routes on the same tunnel. Prefer one CIDR route per subnet over per-host /32 routes where a broad subnet covers all app segment IPs.
  • ZPA bypass means split-tunnel bypass in Cloudflare, not an Access bypass decision. Bypass rules map to WARP Split Tunnel exclude entries. This is a manual configuration step with no API automation - the customer must add bypassed domains and IPs to the device profile split tunnel exclude list through the dashboard.
  • Agentless/browser apps may become separate public-hostname Access apps per domain. WARP private apps remain private-destination apps.
  • The default Cloudflare Access application destination limit is 5 hostnames per app. For ZPA migrations with large app segments, contact the Cloudflare account team to request an increase (up to 50) before implementation. Confirm the limit is active on the account before creating apps - without it, large segments must be split into multiple apps with identical policies, significantly increasing object count.
  • IP-anchored apps require an explicit egress decision before migration: preserve source IP through customer egress, use Cloudflare dedicated egress where available, or accept that the target service must be updated to allow new source IPs. This is a customer decision that blocks implementation if unresolved.
  • Resolver policies can be account-wide. Be careful with overlapping private DNS namespaces across sites or virtual networks; retrieve resolver policy docs before making DNS changes.
  • Each ZPA access policy rule maps to a Cloudflare reusable Access policy. Create all reusable policies before attaching them to Access apps. In default-deny Gateway Network environments, additionally create a Network allow rule with selector "Self-hosted Access App with Private Address is Present" (wirefilter: any(access.private_app[*] in {"*"})) at higher precedence than any broad L4 block rules - without it, Gateway blocks private app traffic before Access policy evaluation occurs.
  • In combined ZIA and ZPA migrations, Gateway Network rules can accidentally block Access private-app traffic. The Gateway Network allow rule above is the fix - place it at higher precedence (lower number) than ZIA-migrated block rules. Add and validate this rule before enabling broad L4 blocks.

Palo Alto / Prisma / NGFW

  • One Palo Alto rule can produce multiple Cloudflare resources. Preserve rule intent, not rule count.
  • App-ID, URL category, zone, HIP, schedule, and decryption behavior rarely translate exactly. Mark partial mappings rather than forcing false equivalence.
  • Export address/service objects and groups with rules. Missing object exports cause silent-looking drops unless explicitly detected.
  • Broad any destination/service rules and very broad CIDRs require manual review. Do not auto-create broad catchalls.
  • HIP/device checks require Cloudflare device posture integrations before enforcement.

Gotchas

  • Source exports often split references across files. Resolve IDs against object, service, and group files before declaring a rule unmappable.
  • Individual users, local groups, departments, and dynamic application IDs often need identity normalization. SCIM/group sync is the gating prerequisite for group selectors.
  • Zscaler caution/warn behavior, Palo Alto App-ID behavior, and TLS/decryption exceptions may not have exact equivalents. Flag them as decision points instead of forcing a 1:1 mapping.
  • Preserve source rule order and hit counts where available. Disable or delete stale/no-hit rules only with user approval.
  • Never create broad allow-all catchalls to preserve connectivity unless explicitly requested and time-limited.

Validation Gates

  • After each migration stage, compare Cloudflare object counts against parsed source counts. Stop on mismatches.
  • Review every unsupported, partial, unmapped, needs_identity, needs_posture, and manual_review item before enabling policies.
  • Validate group matching with real pilot users after SCIM sync and re-authentication.
  • Test TLS inspection and Do Not Inspect behavior before enabling HTTP/DLP blocks broadly.
  • Keep rollback paths explicit: disable migrated rules by prefix, restore source routing, or revert the pilot group/site.
  • Before declaring done, produce a source-rule accounting table: migrated object, partial mapping, not migrated reason, security impact, and owner for each manual action.

Assessment Template

## Migration Assessment

Source stack:
Artifacts reviewed:
Assumptions / missing exports:
Recommended Cloudflare One target:
Mapping summary:
Risks / partial mappings:
Not migrated:
Pilot plan:
Validation:
Rollback:

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Expert in 2000s-era music visualization (Milkdrop, AVS, Geiss) and modern WebGL implementations. Specializes in Butterchurn integration, Web Audio API AnalyserNode FFT data, GLSL shaders for audio-reactive visuals, and psychedelic generative art. Activate on "Milkdrop", "music visualization", "WebGL visualizer", "Butterchurn", "audio reactive", "FFT visualization", "spectrum analyzer". NOT for simple bar charts/waveforms (use basic canvas), video editing, or non-audio visuals.

日本語の概要は準備中です。原文の説明を表示しています。

curiositech/port-daddy22026年10月12日 更新

Expert legal research agent for finding and scraping expungement data state by state. Knows authoritative sources, URL patterns, Firecrawl configuration, and 2026 legal landscape.

日本語の概要は準備中です。原文の説明を表示しています。

curiositech/port-daddy22026年10月12日 更新

Expert in 3D computer vision labeling tools, workflows, and AI-assisted annotation for LiDAR, point clouds, and sensor fusion. Covers SAM4D/Point-SAM, human-in-the-loop architectures, and vertical-specific training strategies. Activate on '3D labeling', 'point cloud annotation', 'LiDAR labeling', 'SAM 3D', 'SAM4D', 'sensor fusion annotation', '3D bounding box', 'semantic segmentation point cloud'. NOT for 2D image labeling (use clip-aware-embeddings), general ML training (use ml-engineer), video annotation without 3D (use computer-vision-pipeline), or VLM prompt engineering (use prompt-engineer).

日本語の概要は準備中です。原文の説明を表示しています。

curiositech/port-daddy22026年10月12日 更新

Apply crisis decision-making research to agent routing, uncertainty triage, and coordination failure analysis in time-pressured systems. Use when diagnosing handoff failures, analytical paralysis, or expert judgment under incomplete information. NOT for routine coding, simple CRUD design, or static single-agent tasks with complete information.

日本語の概要は準備中です。原文の説明を表示しています。

curiositech/port-daddy22026年10月12日 更新

Use for insight, reframing, contradiction, impasse, and anomaly-driven problem solving when execution effort no longer helps. NOT for routine optimization, error correction, or well-specified tasks with known solution paths.

日本語の概要は準備中です。原文の説明を表示しています。

curiositech/port-daddy22026年10月12日 更新

Apply cognitive task analysis to expert work that depends on perceptual cues, branching judgment, and recurring monitoring loops. Use when decomposing expert capability into agent structure, simulation design, or validation interviews. NOT for ordinary step-by-step SOP capture or simple pipelines with no tacit cue layer.

日本語の概要は準備中です。原文の説明を表示しています。

curiositech/port-daddy22026年10月12日 更新

curiositech のスキルをすべて見る

このスキルの問題を報告する