App Store Preflight
Scan iOS/macOS projects for App Store rejection patterns before submission.
Evidence shows: The skill uses a CLI tool (asc) to pull metadata and scans local project files, so permissions are relatively limited (least_privilege=1). However, no user confirmation mechanism is provided (user_confirmation=0). Data flow transparency is partially addressed (data_flow_transparency=1) but does not clarify how data is processed or stored. Sensitive data handling (sensitive_data_handling=1) only mentions privacy manifest checks, not actual data processing. Dependency security (dependency_security=0) lacks dependency audits or vulnerability checks. External effects (external_effects=1) involve modifying project files but no rollback mechanism is described (rollback=0). Source attribution (source_attribution=1) is only via README and LICENSE; publisher identity is unverified.
Evidence shows: Rules and guideline files are consistently structured, indicating good self-consistency (self_consistency=2). Dependency availability (dependency_availability=1) relies on external CLI and skills without version pinning or mirrors. Failure messages (failure_messages=1) lack explicit error handling or user guidance.
Evidence shows: Targets iOS/macOS developers and covers multiple app types (audience_and_scenarios=2). Capability boundaries (capability_boundaries=2) are clear through rule categories and guideline index. Trigger precision (trigger_precision=2) is defined via rule descriptions and detection methods. Environment fit (environment_fit=2) supports macOS/Linux but not Windows.
Evidence shows: Information architecture (information_architecture=2) is clear with directory structure and README. Install notes (install_notes=2) provide npx command. Naming stability (naming_stability=1) lacks versioning or naming conventions. Examples and FAQ (examples_and_faq=1) have examples but no FAQ. Known limitations (known_limitations=1) are not explicitly listed. License (license=2) is MIT. Versioning and changelog (versioning_changelog=0) are missing. Maintenance responsibility (maintenance_responsibility=1) is only via GitHub repository without explicit maintainer.
Evidence shows: Output usability (output_usability=2) provides severity levels and resolution steps. Marginal value (marginal_value=2) offers automated checks saving time. Cost-benefit (cost_benefit=2) is free and open-source but depends on external tools.
Evidence shows: Claim traceability (claim_traceability=2) rules reference specific guideline numbers. Cross-source corroboration (cross_source_corroboration=1) relies solely on Apple guidelines without cross-validation. Fact-inference separation (fact_inference_separation=1) mixes facts and inferences in rules.
- Publisher identity is unverified; use with caution.
- Relies on external CLI and skills without version pinning; may break due to upstream changes.
- No user confirmation mechanism; back up before automatic modifications.
- No rollback mechanism; changes are hard to revert.
- No versioning or changelog; updates are hard to track.
What does this agent do, and when should you use it?
App Store Preflight is an AI agent skill that runs pre-submission checks on iOS/macOS projects to catch common mistakes that lead to App Store rejection. It scans Xcode project files, source code, metadata, and configuration files, flagging issues against Apple's review guidelines. The skill integrates with the `asc` CLI (App Store Connect CLI) and the ASC CLI Skills to pull and inspect App Store metadata. The repository includes a complete index of 100+ Apple Review Guidelines and 10 app-type-specific checklists, organized under `references/guidelines/`. Rejection rules are categorized in `references/rules/` covering metadata, subscriptions, privacy, design, and entitlements. After scanning, it reports findings with severity, affected files, and resolution steps, and supports autofix and revalidation. Installation is via `npx skills add`, and the skill is MIT-licensed.
The skill performs pre-submission checks on iOS/macOS projects. First, it identifies the app type (e.g., subscriptions, social, kids) and loads the corresponding checklist from references/guidelines/by-app-type/. Then it pulls metadata using asc metadata pull --app "<APP_ID>" --version "<VERSION>" --dir ./metadata or via the asc-metadata-sync skill. It scans the project against rejection rules in references/rules/, which are organized by category (e.g., metadata/, subscription/, privacy/, design/, entitlements/). For each rule, it checks for specific patterns such as incomplete review notes, competitor terms, Apple trademark misuse, missing privacy manifests, or unused entitlements. It then reports findings with severity, affected files, and resolution steps, and can apply fixes and re-run affected checks. The skill is driven by instructions in SKILL.md for AI agents.
- An independent iOS developer before submitting a new app to App Store Connect, ensuring review notes are complete.
- A team with subscription-based apps verifying that pricing is not misleading and terms/policies are properly linked.
- Developers of social apps with user-generated content checking compliance with relevant guidelines.
- macOS app developers ensuring their app meets Mac App Store specific requirements.
- Developers using AI features ensuring the app aligns with AI app guidelines (e.g., content generation, privacy).
- Teams using App Store Connect CLI to integrate preflight checks into their CI/CD pipeline.
What are this agent's strengths and limitations?
- Comprehensive coverage with 100+ guidelines and 10 app-type checklists.
- Modular rule structure allows easy addition or customization of rules.
- Integrates with App Store Connect CLI for real metadata pulls.
- Includes autofix and revalidation capabilities to streamline the process.
- MIT licensed, so it can be freely used and modified.
- Dependency on external
ascCLI and ASC CLI Skills incurs additional setup. - Metadata examples assume canonical JSON from
asc; fastlane metadata requires adaptation. - Some rules (e.g., privacy manifest) may require manual Xcode configuration.
- The skill relies on AI agents following
SKILL.md; it is not a standalone executable. - No pre-built CI integration or test suite provided, requiring custom setup.
How do you install or deploy this agent?
Install via npm: Run npx skills add truongduy2611/app-store-preflight-skills. Requires Node.js and the asc CLI (install via brew install asc).
How do you use this agent?
After installation, refer to SKILL.md for full AI agent instructions. Typical workflow: 1) Load the checklist for your app type; 2) Pull metadata using asc metadata pull --app "<APP_ID>" --version "<VERSION>" --dir ./metadata or via the asc-metadata-sync skill; 3) Have the agent scan the project and generate a report; 4) Apply fixes as suggested and re-run checks. Ensure you have App Store Connect API credentials configured.
FAQ
Is this skill free to use?
asc CLI (also free) and may require App Store Connect API credentials.What permissions do I need for the asc CLI?
Can I use this with fastlane metadata?
asc metadata pull. You can either adapt path examples or pull the canonical layout first.Does this guarantee my app will be approved?
Can I add custom rules?
references/rules/ subdirectory following the format in the README.