From movies-for-ai-agents
Browse what's popular, trending, newly released, currently in theaters, or filter by IMDb rating / interest / director. Use when the user asks "what's popular right now", "what's trending", "what's new", "what's in theaters", "show me this week's movies", "anything good out lately", "sci-fi with IMDb 8+", "heist movies", "Nolan films", "hidden gems", or wants to see a feed of current hot titles. Not for personalized picks — for that, use `recommend`. Not for finding a specific title — for that, use `find-movie`.
How this skill is triggered — by the user, by Claude, or both
Slash command
/movies-for-ai-agents:discover-moviesThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are surfacing a feed of notable movies. Two data sources are wired in — route per intent.
You are surfacing a feed of notable movies. Two data sources are wired in — route per intent.
movies_imdb_discover (imdbapi.dev) when the user's query involves…minAggregateRating.movies_imdb_find_name({ query }) to resolve the nameId, then movies_imdb_discover({ nameIds: [...] }).movies_imdb_interests to find the right interestId (in...), then filter on it.minAggregateRating: 7.5, maxVoteCount: 50000, sortBy: "SORT_BY_USER_RATING".languageCodes or countryCodes.movies_trending with window: "week" (or "day"). IMDb's popularity sort is lifetime, not momentum.movies_popular.movies_now_playing. Pass region if the user mentioned a country.movies_trending { window: "week" } — it weights recency.Decide the route. Call the chosen tool.
Filter against user taste (always). Call preferences_get in parallel. Drop results whose genres intersect dislikedGenres. Flag entries already in watched_list as [seen] — don't remove them.
Present 5–8 picks in a readable format. Default to bold + IMDb hyperlink for the title; IMDb rating is the score users trust:
**[Title](https://www.imdb.com/title/{imdbId}/)** (Year[, Country]) — IMDb 8.2 — 128 min
[Genre, Genre] · [Interest, Interest]
One-sentence hook from plot.
Append [seen] if the entry is already in watched_list.
TMDB list results (movies_trending, movies_popular, movies_now_playing, movies_discover) don't carry IMDb data. For the top 3–5 picks, call movies_details({ movieId }) to hydrate imdbId + merged IMDb rating — those entries get the hyperlink. The remaining entries render as plain **Title** with TMDB 7.9 as the rating fallback.
❌ Never use a TMDB URL (themoviedb.org/movie/...) as a fallback link. Plain bold **Title** is the ONLY acceptable degraded state when imdbId is unavailable. Never dual-list both scores unless the user explicitly asks for the TMDB score.
Offer follow-ups. "Want trailer / full details?" → movies_details (TMDB, merges IMDb signal). "Add to watchlist?" → watchlist_add. "Add X to my [name] list?" → list_add({ listName, movieId }) — auto-creates the list. Confirm in one sentence: "Added Weapons to halloween (4 total)." If the user names a list that doesn't exist yet, just create it — no need to ask.
movies_imdb_discover({ genres: ["Sci-Fi"], minAggregateRating: 8, startYear: 2010, endYear: 2019, sortBy: "SORT_BY_USER_RATING", sortOrder: "DESC" })movies_imdb_interests, then movies_imdb_discover({ interestIds: ["in..."], sortBy: "SORT_BY_USER_RATING", sortOrder: "DESC", minVoteCount: 50000 })movies_imdb_find_name({ query: "Christopher Nolan" }) → movies_imdb_discover({ nameIds: ["nm0634240"], sortBy: "SORT_BY_YEAR", sortOrder: "DESC" })movies_imdb_discover({ minAggregateRating: 7.5, maxVoteCount: 50000, sortBy: "SORT_BY_USER_RATING", sortOrder: "DESC", limit: 10 })movies_trending({ window: "week" }) — do NOT route to imdbapi.dev.movies_imdb_discover({ genres: ["Thriller"], countryCodes: ["KR"], minAggregateRating: 7, sortBy: "SORT_BY_USER_RATING", sortOrder: "DESC" })movies_now_playing — the data is region-dependent.movies_imdb_discover returns nothing (over-filtered), relax one constraint at a time (votes, year, rating) before giving up.docs/api-reference/imdbapi-dev.md.npx claudepluginhub jcodesmore/jcodesmore-plugins --plugin movies-for-ai-agentsSearches 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.