aso

Files (1)

• SKILL.md

SKILL.md (first lines)

# ASO — App Store Optimization Skill

## Consolidated Skill Routing

Use five public skills:

- `aso`: full audit, orchestration, and strategy. For roadmap detail, read
  `../skills/aso-plan/GUIDE.md`.
- `aso-listing`: keywords, metadata, localization, and seasonal changes.
- `aso-creative`: screenshots, video, conversion, and experiments.
- `aso-market`: reviews and competitors.
- `aso-release`: technical readiness, compliance, and launch.

The former specialist files are internal `GUIDE.md` references. They preserve
the detailed methods but do not appear as separate skills.

Comprehensive ASO analysis for iOS App Store and Google Play apps. Works in two modes:

1. **Local mode** (default): Run `/aso audit` inside an app project — auto-detects Fastlane metadata, Xcode project, or Gradle project and audits your pre-submission metadata.
2. **Remote mode**: Run `/aso audit <app-id>` with an app ID or store URL — fetches and audits the live listing. Also works for competitor apps.

## Quick Reference

| Command | What it does |
|---------|-------------|
| `/aso audit` | Auto-detect local project and audit its metadata |
| `/aso audit <app-id>` | Audit a live listing by app ID or store URL |
| `/aso audit --compare <app-id>` | Audit local metadata AND compare against live listing |
| `/aso keywords <seeds>` | Keyword research — difficulty, volume, placement strategy |
| `/aso metadata` | Optimize local metadata (or `/aso metadata <app-id>` for remote) |
| `/aso visuals` | Analyze local screenshots (or `/aso visuals <app-id>` for remote) |
| `/aso reviews <app-id>` | Reviews — sentiment analysis, response templates, keyword extraction |
| `/aso competitors <app-id>` | Competitor analysis — keyword gaps, metadata comparison |
| `/aso localization` | Audit local locale coverage (or `/aso localization <app-id>`) |
| `/aso ab-testing <app-id>` | A/B testing — iOS PPO / Android experiment design |
| `/aso technical <app-id>` | Technical ASO — Android Vitals, app size, crashes, update cadence |
| `/aso conversion <app-id>` | Conversion — install rate benchmarking, first-impression audit |
| `/aso compliance` | Store policy check — validates metadata against Apple/Google guidelines |
| `/aso plan <category>` | Strategic plan — 4-phase roadmap with category templates |
| `/aso launch` | Launch — pre-launch checklist using local project state |
| `/aso seasonal <app-id>` | Seasonal — trending keyword opportunities, holiday calendar |

## Auto-Detection (Local Mode)

When no app ID is provided, run project detection first:
```bash
uv run python scripts/detect_project.py --json
```

This scans the current directory for:

| Source | What it detects | Files checked |
|--------|----------------|---------------|
| **Fastlane iOS** | Store metadata per locale (name, subtitle, keywords, description) | `fastlane/metadata/{locale}/name.txt`, `subtitle.txt`, `keywords.txt`, etc. |
| **Fastlane Android** | Store metadata per locale (title, short/full description) | `fastlane/metadata/android/{locale}/title.txt`, `full_description.txt`, etc. |
| **Xcode project** | Bundle ID, app name | `*.xcodeproj/project.pbxproj`, `Info.plist` |
| **Gradle project** | Application ID, app name | `app/build.gradle`, `AndroidManifest.xml`, `res/values/strings.xml` |

Priority: Fastlane metadata (richest data) > Xcode/Gradle (basic identifiers only).

If Fastlane metadata is found, all fields are available for validation, keyword analysis, and optimization. If only Xcode/Gradle is found, extract the app ID and optionally fetch the live listing for a full audit.

## Remote Mode

When an app ID or URL is provided, fetch the live listing:

- **iOS**: Numeric ID (`id284882215`), Apple App Store URL (`apps.apple.com/...`)
- **Android**: Package name (`com.whatsapp`), Google Play URL (`play.google.com/store/apps/details?id=...`)

```
iOS:     id284882215  OR  https://apps.apple.com/app/id284882215
Android: com.whatsapp  OR  https://play.google.com/store/apps/details?id=com.whatsapp
```

## Compare Mode

`/aso audit --compare <app-id>` audits BOTH local metadata and the live listing, then shows:
- Fields that differ between local and live
- Whether local changes would improve or regress the score
- Pre-submission validation of pending metadata changes

## Category Detection

After fetching listing data, classify the app:
- **Gaming**: game category, IAP patterns, leaderboards
- **SaaS/Productivity**: subscription model, enterprise features, workspace tools
- **Health/Fitness**: HealthKit/Google Fit, workout terms, wellness
- **E-commerce**: shopping category, payment, cart, product catalog
- **Social**: messaging, feed, connections, sharing
- **Fintech**: banking, payments, investment, crypto
- **Other**: education, travel, utilities, news, entertainment

Category informs which plan templates and benchmarks to apply.

## Audit Orchestration Flow

When running `/aso audit <app-id>`:

### Step 1 — Data Fetch (sequential)
```bash
uv run python scripts/fetch_listing.py <app-id> --country us --json
```
Parse the result to get structured listing data.

### Step 2 — Category Detection (sequential)
Classify app from listing metadata. Load category-specific benchmarks.

### Step 3 — Parallel Subagent Dispatch
Run all applicable specialists in parallel when the harness supports
delegation. Otherwise, load the corresponding specialist skills and run them
sequentially. Do not continue to scoring until all applicable results return.

| Agent | Focus |
|-------|-------|
| `aso-keywords` | Keyword coverage in current metadata |
| `aso-metadata` | Field quality, char limits, keyword placement |
| `aso-visuals` | Screenshot count, narrative, icon, video |
| `aso-reviews` | Sentiment distribution, themes, response rate |
| `aso-competitors` | Top 5 competitor comparison |

From FelixGraeber/claude-aso-audit-skill, MIT.

More from FelixGraeber/claude-aso-audit-skill