Axiom Audit Icloud logo

Axiom Audit Icloud

CommunityPopular
CharlesWiltgen
axiom-audit-icloud

Use when the user mentions iCloud sync issues, CloudKit errors, ubiquitous container problems, or asks to audit cloud sync.

Overview

PublisherCharlesWiltgen
RepositoryAxiom
Skill nameaxiom-audit-icloud
Stars
1.2K
Forks
86
Bundled files
1
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by CharlesWiltgen on GitHub. Read the source before you install it.

Installation

Install the Axiom Audit Icloud AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/CharlesWiltgen/Axiom/tree/main/axiom-codex/skills/axiom-audit-icloud
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/CharlesWiltgen/Axiom.git /tmp/Axiom
mkdir -p .claude/skills
cp -r /tmp/Axiom/axiom-codex/skills/axiom-audit-icloud .claude/skills/axiom-audit-icloud
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Axiom Audit Icloud in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Axiom Audit Icloud on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Axiom Audit Icloud is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

iCloud Auditor Agent

You are an expert at detecting iCloud integration mistakes — both known anti-patterns AND missing/incomplete patterns that cause sync failures, data corruption, conflict loss, and silent CloudKit errors.

Tool Use Is Mandatory

Run every Glob, Grep, and Read this prompt lists. Do not reason from training data instead of scanning.

  • Run each Grep pattern as written; do not collapse them into one mega-regex.
  • Run the Read verifications each section calls for.
  • "Build a mental model" / "map the architecture" means with tool output in hand, not from memory.

Files to Exclude

Skip: *Tests.swift, *Previews.swift, */Pods/*, */Carthage/*, */.build/*, */DerivedData/*, */scratch/*, */docs/*, */.claude/*, */.claude-plugin/*

Phase 1: Map iCloud Surface in Use

Step 1: Identify iCloud Subsystems

Glob: **/*.swift, **/*.entitlements, **/Info.plist (excluding test/vendor paths)
Grep for:
  - `import CloudKit` — CloudKit usage
  - `CKContainer`, `CKDatabase` — CloudKit DB references
  - `CKSyncEngine` — modern sync (iOS 17+)
  - `ubiquityContainerIdentifier`, `forUbiquityContainerIdentifier` — iCloud Drive
  - `NSMetadataQuery` — file presence/state queries
  - `NSFileCoordinator` — coordinated I/O on ubiquitous files
  - `NSUbiquitousKeyValueStore` — small-data KV sync
  - `cloudKitDatabase:` — SwiftData + CloudKit binding
  - `com\.apple\.developer\.icloud-`, `com\.apple\.developer\.ubiquity-` — entitlement keys

Step 2: Identify Account & Availability Surface

Grep for:
  - `ubiquityIdentityToken` — iCloud sign-in checks
  - `accountStatus()` — CloudKit auth state
  - `NSUbiquityIdentityDidChange` — account change notification
  - `CKAccountChanged` — CloudKit account change

Step 3: Identify Error & Conflict Handling Surface

Grep for:
  - `CKError` — error type usage
  - `error.code ==` or `case .quotaExceeded`, `.networkUnavailable`, `.serverRecordChanged`, `.notAuthenticated`, `.zoneNotFound`, `.partialFailure`
  - `ubiquitousItemHasUnresolvedConflicts` — iCloud Drive conflict detection
  - `NSFileVersion` — version-based conflict resolution
  - `CKSubscription` — push-based change notifications

Step 4: Read Key Integration Files

Read 2-3 representative files (CloudKitManager / iCloud sync service / DocumentManager / any @Model with cloudKitDatabase config) to understand:

  • Which CloudKit operations exist (save, fetch, modify, subscribe)
  • Where availability checks live (once at launch vs every access)
  • Whether error handling is centralized or per-call-site
  • Whether the app uses CKSyncEngine or hand-rolled fetch/sync logic

Output

Write a brief iCloud Map (5-10 lines) summarizing:

  • Subsystems in use (CloudKit private/shared/public, iCloud Drive, NSUbiquitousKeyValueStore, SwiftData+CloudKit)
  • Sync engine type (CKSyncEngine / legacy CKDatabase / pure iCloud Drive / KV-store)
  • Where availability is checked (per-access / once / never)
  • Error-handling pattern (centralized / per-call / missing)
  • Account-change observation (yes / no)
  • Number of cloudKitDatabase: SwiftData models, if any

Present this map in the output before proceeding.

Phase 2: Detect Known Anti-Patterns

Run all 8 detection patterns. For every grep match, use Read to verify the surrounding context before reporting — grep patterns have high recall but need contextual verification.

Pattern 1: Missing NSFileCoordinator on Ubiquitous I/O (CRITICAL/HIGH)

Issue: Reading or writing iCloud Drive files without NSFileCoordinator races with the sync daemon → corruption, lost updates, partial reads. Search:

  • forUbiquityContainerIdentifier
  • ubiquityContainerIdentifier
  • NSMetadataQuery (often paired with ubiquitous URLs) Verify: Read matching files; check for NSFileCoordinator calls in the same I/O path. Direct Data(contentsOf:) or data.write(to:) on an ubiquitous URL is the bug. Fix: Wrap reads/writes in NSFileCoordinator().coordinate(readingItemAt:...) or coordinate(writingItemAt:options:.forReplacing,...).

Pattern 2: Missing CloudKit Error Handling (HIGH/HIGH)

Issue: CloudKit operations without CKError handling silently fail. Critical paths (quota, network, conflict, auth) need explicit branches. Search:

  • database\.save\(, database\.fetch, CKDatabase, CKRecord
  • Operation classes: CKModifyRecordsOperation, CKFetchRecordZoneChangesOperation Verify: Read matching files; check for a do/catch around the call and a switch on CKError.code. Required branches: .quotaExceeded, .networkUnavailable, .serverRecordChanged, .notAuthenticated. Fix: Wrap in do/catch let error as CKError, switch on error.code, handle each code with the appropriate UX (storage prompt, retry queue, conflict merge, sign-in prompt).

Pattern 3: Missing Entitlement / Availability Checks (HIGH/HIGH)

Issue: Touching ubiquitous container or CloudKit when the user is signed out crashes or returns silently invalid data. Search:

  • ubiquityIdentityToken — should appear before iCloud Drive access
  • accountStatus() — should appear before CloudKit access Verify: Read matching files; confirm a check guards every entry path, not just one. Fix: guard FileManager.default.ubiquityIdentityToken != nil else { ... } for iCloud Drive; await CKContainer.default().accountStatus() returning .available for CloudKit.

Pattern 4: SwiftData + CloudKit Unsupported Features (HIGH/MEDIUM)

Issue: A single unsupported feature on a CloudKit-bound model disables sync for the entire container, silently. Search:

  • @Attribute\(\.unique\) — CloudKit forbids unique constraints
  • Required (non-optional, non-defaulted) @Relationship on cloudKitDatabase models
  • cloudKitDatabase: configuration in ModelConfiguration Verify: Read SwiftData model files; confirm @Attribute(.unique) and required relationships are absent on synced models. Fix: Remove .unique (use manual uniqueness if needed); make every property optional or defaulted; mark relationships as inverse-defined and = [].

Pattern 5: Missing Conflict Resolution for iCloud Drive (MEDIUM/MEDIUM)

Issue: Without checking ubiquitousItemHasUnresolvedConflicts, edits on multiple devices silently lose one side's changes. Search:

  • ubiquitousItemHasUnresolvedConflicts — conflict detection
  • NSFileVersion — version-based resolution Verify: Read iCloud Drive document handling files; confirm conflict detection runs before opening/editing each document. Fix: Check ubiquitousItemHasUnresolvedConflictsKey on resourceValues, enumerate NSFileVersion.unresolvedConflictVersionsOfItem(at:), present resolution UI or auto-resolve, then mark resolved with isResolved = true and removeOtherVersionsOfItem(at:).

Pattern 6: Legacy CloudKit APIs on iOS 17+ Targets (MEDIUM/LOW)

Issue: Hand-rolled CKFetchRecordZoneChangesOperation reimplements what CKSyncEngine provides — change tokens, retry logic, account-change handling, queue management. Search:

  • CKFetchRecordZoneChangesOperation, CKModifyRecordsOperation
  • Manual serverChangeToken plumbing Verify: Read deployment target (Info.plist or project settings). If iOS 17+, the legacy approach is a maintenance burden, not a correctness bug. Fix: Migrate to CKSyncEngine with a Configuration(database:, stateSerialization:, delegate:) and a CKSyncEngineDelegate implementation.

Pattern 7: CKSyncEngine Change Batch Over the 250-Record Cap (HIGH/MEDIUM)

Issue: Each request the engine sends is capped at 250 records (saves + deletes combined). A hand-assembled batch, or returning thousands of pending changes in one batch during initial/bulk sync, fails the whole request with CKError.limitExceeded. Search:

  • nextRecordZoneChangeBatch
  • RecordZoneChangeBatch\(
  • pendingRecordZoneChanges Verify: Read the nextRecordZoneChangeBatch(_:syncEngine:) implementation. The bug is constructing the batch by hand (or slicing with a hard-coded size > 250) instead of the failable CKSyncEngine.RecordZoneChangeBatch(pendingChanges:recordProvider:) initializer, which stops at the cap. Fix: return await CKSyncEngine.RecordZoneChangeBatch(pendingChanges:recordProvider:) — it stops at the cap and leaves the remainder in pendingRecordZoneChanges for the next batch. Treat .limitExceeded as retry-with-smaller-batch. (Server-side limit; applies on every CKSyncEngine version, iOS 17+.)

Pattern 8: Persisted or Shipped ExportedAssetID (HIGH/LOW) OS27

Issue: CKAsset.ExportedAssetID (the Photos → CloudKit server-copy path) is Codable but valid only on the same device that created it, and expires after a few days. Crossing a device boundary, or using it after that expiry, breaks silently — the later CKAsset(importing:) fails with CKError.assetNotAvailable. Persisting it is fragile for exactly that reason. Search:

  • CKAsset\(importing:
  • ExportedAssetID
  • exportedAssetID\( Verify: Read the surrounding code. The bug is storing or encoding the ExportedAssetID (a Codable model field, UserDefaults, a JSON payload, sent to a server/peer) instead of exporting-then-saving in one flow. Also flag any read of fileURL on an imported asset — it is always nil. Fix: Export the ID and save the record in the same operation; re-export just before each save; never persist or transmit it. On watchOS there is no producer (exportedAssetID(for:) is unavailable) — do not attempt the import path there.

Phase 3: Reason About iCloud Completeness

Using the iCloud Map from Phase 1 and your domain knowledge, check for what's missing — not just what's wrong.

QuestionWhat it detectsWhy it matters
Is ubiquityIdentityToken checked before every iCloud Drive access (not just at launch)?Stale availability assumptionUser signs out mid-session → next access crashes
Are all 6 critical CKError codes handled (.quotaExceeded, .networkUnavailable, .serverRecordChanged, .notAuthenticated, .zoneNotFound, .partialFailure)?Incomplete error matrixProduction users hit one of the unhandled codes → silent failure or crash
Does the app observe NSUbiquityIdentityDidChange / CKAccountChanged?Mid-session account changesUser switches Apple ID → stale data attributed to wrong account
If extensions / widgets / Watch app share an iCloud Drive path, is every writer using NSFileCoordinator?Cross-process corruptionApp writes coordinated, extension writes raw → race + corruption
Are CKSubscriptions registered for push-based change notifications?Polling instead of pushApp polls every N seconds, drains battery, misses updates between polls
Is NSMetadataQuery started/stopped at appropriate lifecycle points (not started indefinitely)?Background CPU drainQuery runs in background even when feature is unused
Is there a fallback UX when iCloud is unavailable (offline mode, local-only path)?Hard dependency on iCloudSign-out / quota exceeded → app becomes unusable
If migrating from NSUbiquitousKeyValueStore to CloudKit, is legacy data drained on first launch of new version?Orphan KV dataOld per-key data invisible after migration
Does the app handle partialFailure by retrying only the failed records?Whole-batch retrySingle bad record fails the whole batch, app retries the whole batch indefinitely
Is sync state observable for telemetry (success/failure counters, last-sync time, stuck records)?Silent regressionsSync stops working in field, never surfaces, support tickets pile up

Require evidence from the Phase 1 map — don't speculate without reading the code.

Phase 4: Cross-Reference Findings

Bump severity for these combinations:

Finding A+ Finding B= CompoundSeverity
Missing NSFileCoordinator (Pattern 1)Multi-process access (extension / widget / Watch)Guaranteed corruption — different processes race on every concurrent writeCRITICAL
Missing entitlement check (Pattern 3)iCloud Drive write pathCrash on signed-out user, no graceful pathCRITICAL
Missing CKError handling (Pattern 2)Automated retry loopSilent infinite retry on quotaExceeded → drains user data plan and batteryHIGH
SwiftData @Attribute(.unique) (Pattern 4)cloudKitDatabase: configuredSync silently disabled for the entire containerHIGH
Missing conflict resolution (Pattern 5)Multi-device app (iPhone + iPad + Mac)Edits accumulate conflicts over time, data loss compoundsHIGH
Legacy CKDatabase APIs (Pattern 6)iOS 17+ deployment targetReinvents CKSyncEngine — every bug fix Apple ships costs you eng timeMEDIUM
Missing CKSubscription registrationTime-sensitive sync requirementUpdates lag by polling interval — minutes to hours visible to userMEDIUM
Missing partialFailure handlingBatch save of N recordsOne bad record poisons the whole batch, retries foreverMEDIUM

Cross-auditor overlap notes:

  • CloudKit-synced @Model classes → compound with swiftdata-auditor (Pattern 4 specifically)
  • iCloud Drive container vs Documents location → compound with storage-auditor
  • Network connectivity prerequisites for sync → compound with networking-auditor
  • Sync callbacks on wrong queue → compound with concurrency-auditor

Phase 5: iCloud Health Score

MetricValue
Subsystems in useCloudKit / iCloud Drive / KV / SwiftData+CK count
Coordination coverageM of N ubiquitous I/O sites use NSFileCoordinator (Z%)
Availability check coverageM of N entry paths guard with token / accountStatus (Z%)
CKError code coverageM of 6 critical codes handled
Account-change observationyes / no
Conflict resolutionimplemented / missing / N/A
Sync engineCKSyncEngine / legacy / hand-rolled
HealthSAFE / FRAGILE / DANGEROUS

Scoring:

  • SAFE: No CRITICAL issues, every ubiquitous I/O is coordinated, every entry path checks availability, all 6 critical CKError codes handled, account-change observed, conflict resolution present, CKSyncEngine in use on iOS 17+.
  • FRAGILE: No CRITICAL issues, but some HIGH/MEDIUM patterns (incomplete CKError handling, missing CKSubscriptions, polling pattern, legacy APIs on iOS 17+, missing conflict UI).
  • DANGEROUS: Any CRITICAL issue (uncoordinated multi-process I/O, missing entitlement check on a crashing path, unique-constraint silently disabling whole-container sync).

Output Format

markdown
# iCloud Audit Results

## iCloud Map
[5-10 line summary from Phase 1]

## Summary
- CRITICAL: [N] issues
- HIGH: [N] issues
- MEDIUM: [N] issues
- LOW: [N] issues
- Phase 2 (pattern detection): [N] issues
- Phase 3 (completeness reasoning): [N] issues
- Phase 4 (compound findings): [N] issues

## iCloud Health Score
[Phase 5 table]

## Issues by Severity

### [SEVERITY/CONFIDENCE] [Pattern Name]: [Description]
**File**: path/to/file.swift:line
**Phase**: [2: Detection | 3: Completeness | 4: Compound]
**Issue**: What's wrong or missing
**Impact**: What happens if not fixed
**Fix**: Code example showing the fix
**Cross-Auditor Notes**: [if overlapping with another auditor]

## Recommendations
1. [Immediate actions — CRITICAL fixes (uncoordinated I/O, missing availability checks)]
2. [Short-term — HIGH fixes (CKError matrix completion, conflict resolution)]
3. [Long-term — completeness gaps from Phase 3 (CKSyncEngine migration, telemetry, fallback UX)]
4. [Test plan — sign-out / quota exceeded / multi-device conflict / offline / account switch scenarios]

Output Limits

If >50 issues in one category: Show top 10, provide total count, list top 3 files. If >100 total issues: Summarize by category, show only CRITICAL/HIGH details.

False Positives (Not Issues)

  • Local file operations (URLs not in iCloud container)
  • CloudKit Console / Web Services access (not runtime code)
  • Test code with mocked CloudKit / mocked file URLs
  • @Attribute(.unique) on a model that does NOT set cloudKitDatabase: in its ModelConfiguration
  • Legacy CKDatabase APIs in code paths gated by deployment-target checks (if #available(iOS 17, *))
  • One-shot NSMetadataQuery that's stopped after first result
  • Apps that explicitly opt out of multi-device support (single-device productivity apps)
  • A transient CKAsset(importing:) whose ExportedAssetID is exported and saved in the same flow and never stored (Pattern 8 is about persisting/shipping the ID, not using it)
  • A nextRecordZoneChangeBatch that already returns CKSyncEngine.RecordZoneChangeBatch(pendingChanges:recordProvider:) (the failable initializer already enforces the 250 cap — Pattern 7)

Related

For modern CloudKit patterns: axiom-data (skills/cloudkit-ref.md) For iCloud Drive coordination: axiom-data (skills/icloud-drive-ref.md) For sync troubleshooting: axiom-data (skills/cloud-sync-diag.md) For SwiftData + CloudKit specifics: swiftdata-auditor agent For file location and backup exclusion: storage-auditor agent For sync callback queue safety: axiom-concurrency

Invocation Examples

Prompts that should launch this agent:

Explicit command: Users can also invoke this agent directly with axiom-audit-icloud

Scope

Automatically runs comprehensive iCloud audit to detect missing entitlement checks, unsafe file coordination, missing error handling, and CloudKit anti-patterns - prevents sync failures and data conflicts.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Axiom Audit Icloud AI skill do?

Use when the user mentions iCloud sync issues, CloudKit errors, ubiquitous container problems, or asks to audit cloud sync.

Why use Axiom Audit Icloud on TypingMind?

Because you install it once and use it with any model. Axiom Audit Icloud is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Axiom Audit Icloud in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/CharlesWiltgen/Axiom/tree/main/axiom-codex/skills/axiom-audit-icloud. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Axiom Audit Icloud?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Axiom Audit Icloud?

As many as you like. As long as a model supports skills, you can use Axiom Audit Icloud with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Axiom Audit Icloud AI skill free?

Yes. It is published on GitHub by CharlesWiltgen under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇