From asc-release-kit
Generate localized ASC keyword-field recommendations for iOS apps. Use when the user asks for App Store Connect keywords, ASO keyword research, localized ASC keyword sets, or a 100-character keyword field. Focus only on the hidden Keywords field, not title, subtitle, screenshots, descriptions, or other ASC metadata.
How this skill is triggered — by the user, by Claude, or both
Slash command
/asc-release-kit:asc-keywordsThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this skill to produce one or more localized App Store Connect `Keywords` fields. The deliverable is keyword-only ASO output: no subtitle, no title rewrite, no screenshot copy, no marketing description.
Use this skill to produce one or more localized App Store Connect Keywords fields. The deliverable is keyword-only ASO output: no subtitle, no title rewrite, no screenshot copy, no marketing description.
Before keyword research, gather enough context to understand the current app. Do not jump straight from a category label to final keywords.
If a local project or workspace is available, inspect the relevant app files first, such as product docs, README, app entry points, feature screens, localization files, existing App Store metadata, or marketing copy. Keep the pass lightweight and read-only unless the user asks for edits.
Build a short working brief before generating candidates:
Keywords.If the project context is missing or too thin, ask focused questions or request the minimum files/evidence needed. It is acceptable to produce a starter keyword set only after labeling it as context-limited and pending App Store validation.
The only input that cannot be inferred: target locales (e.g. en-US, zh-Hans).
If target locales are missing and were not provided by the caller, ask once and stop.
Everything else — app category, core user job, seed words — infer from the product brief or project files. Do not ask the user for these.
Build candidates from the product brief and project evidence:
For each candidate or autocomplete phrase, classify the competitive shape:
avoid: top results are dominated by large brands, platform giants, or exact-brand intent.keep: results include small or indie apps, lower-review apps, or mixed-quality competitors.priority: a relevant long-tail phrase has weak-looking competition and direct product fit.unclear: insufficient evidence; keep only if it is semantically strong and space-efficient.Bias toward long-tail component words that can combine into multiple useful phrases. For a new indie app, do not spend scarce characters on giant head terms unless they are short, essential, and highly reusable.
Build each localized App Store Connect keyword field under the 100-character limit.
word1,word2,word3.<= 100.home,workout rather than home workout.app, free, iphone, ipad, ios, apple.Return a compact table per locale:
| Locale | Keywords | Count | Notes |
|---|---|---|---|
| en-US | journal,diary,mood,habit,photo,calendar | 40 | starter set; validate autocomplete |
Then include:
Priority phrases: 5-10 phrase combinations the token set is intended to cover.Removed: duplicates, giant-brand terms, filler/platform words, and weak terms.Manual validation checklist: exact App Store searches the user should run before shipping.If final evidence is weak, say so plainly. Do not pretend the set is production-grade when it was built without App Store autocomplete or competitor-result checks.
A good final answer:
asc-keywords.md in the current working directory. Confirm the absolute path in the completion summary.npx claudepluginhub raydeveloperf/app-store-connect-release-kit --plugin asc-release-kitSearches MemPalace before answering questions about past work, people, projects, or prior decisions. Returns verbatim stored content instead of guessing from model memory.
Guides Payload CMS config (payload.config.ts), collections, fields, hooks, access control, APIs. Debugs validation errors, security, relationships, queries, transactions, hook behavior.
Implements vector databases with Pinecone, Weaviate, Qdrant, Milvus, pgvector for semantic search, RAG, recommendations, and similarity systems. Optimizes embeddings, indexing, and hybrid search.