← 返回 Skills 市场
anderskev

Cloudkit Code Review

作者 Kevin Anderson · GitHub ↗ · v1.2.1 · MIT-0
cross-platform ✓ 安全检测通过
173
总下载
0
收藏
1
当前安装
2
版本数
在 OpenClaw 中安装
/install cloudkit-code-review
功能描述
Reviews CloudKit code for container setup, record handling, subscriptions, and sharing patterns. Use when reviewing code with import CloudKit, CKContainer, C...
使用说明 (SKILL.md)

CloudKit Code Review

Quick Reference

Issue Type Reference
CKContainer, databases, zones, entitlements references/container-setup.md
CKRecord, references, assets, batch operations references/records.md
CKSubscription, push notifications, silent sync references/subscriptions.md
CKShare, participants, permissions, acceptance references/sharing.md

Review Checklist

  • Account status checked before private/shared database operations
  • Custom zones used (not default zone) for production data
  • All CloudKit errors handled with retryAfterSeconds respected
  • serverRecordChanged conflicts handled with proper merge logic
  • CKErrorPartialFailure parsed for individual record errors
  • Batch operations used (CKModifyRecordsOperation) not individual saves
  • Large binary data stored as CKAsset (records have 1MB limit)
  • Record keys type-safe (enums) not string literals
  • UI updates dispatched to main thread from callbacks
  • CKAccountChangedNotification observed for account switches
  • Subscriptions have unique IDs to prevent duplicates
  • CKShare uses custom zone (sharing requires custom zones)

When to Load References

  • Reviewing container/database setup or zones -> container-setup.md
  • Reviewing record CRUD or relationships -> records.md
  • Reviewing push notifications or sync triggers -> subscriptions.md
  • Reviewing sharing or collaboration features -> sharing.md

Output Format

Report issues using: [FILE:LINE] ISSUE_TITLE

Examples:

  • [AppDelegate.swift:24] CKContainer not in custom zone
  • [SyncManager.swift:156] Unhandled CKErrorPartialFailure
  • [DataStore.swift:89] Missing retryAfterSeconds backoff

Review Questions

  1. What happens when the user is signed out of iCloud?
  2. Does error handling respect rate limiting (retryAfterSeconds)?
  3. Are conflicts resolved or does data get overwritten silently?
  4. Is the schema deployed to production before App Store release?
  5. Are shared records in custom zones (required for CKShare)?

Hard gates (before reporting)

Complete in order for each finding you intend to report. Do not advance until the pass condition is satisfied.

  1. Location artifact — The finding includes [FILE:LINE] (or a line range) copied from the current file contents; the path resolves in this repo.
  2. Scope read — You read the full surrounding unit: the type or function that owns the CloudKit work (for example the CKOperation subclass usage, completion handler chain, or CKRecord lifecycle), not only a diff hunk or isolated snippet.
  3. CloudKit or deployment claim (only if the finding depends on container identifiers, public vs private database choice, custom zone requirement, iCloud account state, entitlements, or production schema) — You name one concrete artifact you inspected (for example com.apple.developer.icloud-container-environment or container ID in the entitlements file, CKContainer.default() vs custom identifier in source, Info.plist / target capability, or evidence that schema is deployed) or you downgrade the item to an open question in Review Questions.
  4. Protocol — Pre-report steps in review-verification-protocol are satisfied for this item (no finding if they are not).

Use the issue format [FILE:LINE] ISSUE_TITLE for each reported finding. Hard gate 4 is the full pre-report checklist for this skill’s review type.

安全使用建议
This skill is an instruction-only CloudKit code-review checklist and appears internally consistent: it will read your repository files (source files, Info.plist/entitlements) to check CloudKit usage and produce line-referenced findings. There are no installs or secret requests. Two practical cautions: (1) provenance is unknown (no homepage/source listed) — if you need third-party trust, prefer skills from known authors or inspect the SKILL.md and reference docs yourself before enabling; (2) the agent will read repo contents when run, so don’t run it on repositories containing secrets you don’t want reviewed. Otherwise it’s coherent for its stated purpose.
能力评估
Purpose & Capability
Name/description match the files and instructions: the skill is explicitly a CloudKit code reviewer and only includes reference docs and a runtime checklist. It requests no binaries, env vars, or installs that would be unrelated to static code review. (Note: the package source/homepage is unknown, but that affects provenance, not coherence.)
Instruction Scope
SKILL.md instructs the agent to read repository source (files, types, functions, entitlements, Info.plist) and produce file/line-tagged findings — this stays within the stated goal of reviewing CloudKit usage. The 'hard gates' deliberately require reading surrounding units and concrete artifacts; that is appropriate for code review and does not instruct access to unrelated system secrets or external endpoints.
Install Mechanism
No install spec and no code files — instruction-only. Nothing is downloaded or written to disk during install, so install risk is minimal.
Credentials
No environment variables, credentials, or config paths are requested. The review instructions reasonably require inspecting repository files (entitlements, Info.plist) to verify claims about containers and entitlements.
Persistence & Privilege
always:false and default autonomous invocation behavior. The skill does not request persistent system-level changes or cross-skill modifications. Autonomous invocation is allowed by platform default but is not combined with broad credential access or other red flags.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install cloudkit-code-review
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /cloudkit-code-review 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.2.1
- Added a "Hard gates (before reporting)" section outlining mandatory pre-report review steps. - Specified four required gates: location artifact, scope read, CloudKit/deployment claim or downgrade, and protocol verification. - Clarified that findings must follow these steps prior to being reported, reinforcing review discipline. - No changes to core checklist, guidance, or output format.
v1.2.0
cloudkit-code-review v1.2.0 - Added comprehensive skill description and usage guidance in SKILL.md. - Introduced a detailed review checklist for CloudKit best practices. - Provided quick reference links for container setup, record handling, subscriptions, and sharing. - Clarified expected output format when reporting issues. - Included key review questions for evaluating CloudKit code.
元数据
Slug cloudkit-code-review
版本 1.2.1
许可证 MIT-0
累计安装 1
当前安装数 1
历史版本数 2
常见问题

Cloudkit Code Review 是什么?

Reviews CloudKit code for container setup, record handling, subscriptions, and sharing patterns. Use when reviewing code with import CloudKit, CKContainer, C... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 173 次。

如何安装 Cloudkit Code Review?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install cloudkit-code-review」即可一键安装,无需额外配置。

Cloudkit Code Review 是免费的吗?

是的,Cloudkit Code Review 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Cloudkit Code Review 支持哪些平台?

Cloudkit Code Review 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Cloudkit Code Review?

由 Kevin Anderson(@anderskev)开发并维护,当前版本 v1.2.1。

💬 留言讨论