Aso Compliance

ASO Compliance — Store Policy Checker

Validates app metadata against Apple App Store Review Guidelines and Google Play Developer Program Policies. Catches violations before they cause rejection.

Two Modes

1. Standalone: /aso compliance or /aso compliance — check current metadata
2. Audit integration: Runs automatically as part of /aso audit after all agents return, validating both current metadata AND recommended changes

Process

Step 1: Get Metadata

Same as other skills — local detection or remote fetch:

uv run python scripts/detect_project.py --json   # local mode
uv run python scripts/fetch_listing.py  --json  # remote mode

Step 2: Load Policy Reference

Read references/store-policies.md for the complete rule set covering:

• Title restrictions and banned words
• Description content prohibitions
• Keyword spam rules
• Screenshot requirements
• Pricing/IAP disclosure rules
• Health/medical claim restrictions
• Financial claim restrictions
• Review manipulation rules
• AI content policies
• Intellectual property rules

Step 3: Run Compliance Checks

For each metadata field, check against every applicable rule for the detected platform.

Title checks:

• Banned words present? (Google: "#1", "best", "top", "free", "download now", etc.)
• Competitor brand names?
• Trademarked terms without license?
• ALL CAPS? (Google: prohibited unless registered brand)
• Emojis? (Google: prohibited in title)
• Other platform names? (Apple: no "Android")
• Pricing language?

Description checks:

• Testimonials? (Google: prohibited)
• Rating/performance claims? ("4.9 stars", "million downloads")
• Pricing info? ("$X off", "free trial", "limited time")
• Call-to-actions? (Google: "download now", "install now")
• Competitor names?
• Keyword stuffing? (unnatural density)
• Misleading functionality claims?

Category-specific checks:

• Health apps: disclaimers present? No unvalidated measurement claims?
• Finance apps: APR disclosed? Financial features declaration done?
• AI apps: data sharing disclosed? In-app reporting?
• Gaming: loot box odds disclosed? IAP transparent?

IAP/pricing checks:

• If app has IAP: disclosed in description and screenshots?
• If "free": any essential features behind paywall?
• Subscription terms clear?

Screenshot checks (if descriptions/captions available):

• Promotional text present? ("#1", "Best", "Sale")
• Ranking claims?
• Non-app UI elements?

Step 4: Report

Severity levels:

| Level | Meaning | Icon | |-------|---------|------| | VIOLATION | Will likely cause rejection | Must fix | | RISK | May trigger review or penalty | Should fix | | ADVISORY | Best practice recommendation | Consider |

Output format:

## Compliance Report: [App Name]

**Platform**: [iOS / Android / Both]
**Result**: [X violations, Y risks, Z advisories]
**Compliance Score**: XX/100

### Violations (Must Fix)
| # | Field | Rule | Issue | Fix |
|---|-------|------|-------|-----|
| 1 | title | Google Banned Words | Contains "#1" | Remove and use descriptive keyword |

### Risks (Should Fix)
| # | Field | Rule | Issue | Fix |
|---|-------|------|-------|-----|

### Advisories (Consider)
| # | Field | Rule | Issue | Fix |
|---|-------|------|-------|-----|

### Recommendation Flags
[If running as part of audit — list any agent recommendations that violate policies]
| Source | Recommendation | Violation | Alternative |
|--------|---------------|-----------|-------------|

When Running as Part of Audit

The compliance agent runs AFTER all other agents return (Step 3.5 in audit flow). It receives:

1. Current metadata (same as other agents)
2. Recommendations from all agents (title rewrites, keyword suggestions, description changes)

It validates ALL recommendations against store policies. Any violating recommendation is:

• Flagged in the compliance report
• Removed from the final ASO-ACTION-PLAN.md
• Replaced with a compliant alternative where possible

This ensures the audit NEVER recommends metadata that would get the app rejected.

Available Tools

Read, Bash, Write, Glob, Grep

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: FelixGraeber
• Source: FelixGraeber/claude-aso-audit-skill
• License: MIT

Install and usage instructions live in the source repository linked above.