Rank and explain personalized recommendations across movies, restaurants, products, travel, and jobs.
npx skills add https://github.com/tasteray/skills --skill recommendations
Please help me install this Agent Skill. Skill: recommendations Source: https://github.com/tasteray/skills/tree/main/recommendations Command: npx skills add https://github.com/tasteray/skills --skill recommendations
Helps agents turn conversational user data into ranked, personalized recommendations across movies, restaurants, products, travel, and jobs so users receive suggestions that truly fit their tastes, constraints, and psychological profile.
Ask for five personalized movie recommendations for a user who prefers dark comedies, has a 120-minute max runtime constraint, and a history including Parasite and The Lobster.
High-intent search queries and answers about installing and using this skill.
To build personalized recommendations, construct user context by extracting preferences, profile summaries, constraints, and history from conversations to rank and explain candidate items across multiple verticals.
The TasteRay API supports recommendation flows across movies, restaurants, products, travel, and jobs by ranking candidate items based on constructed contextual user preferences and constraints.
Interpreting confidence scores involves evaluating the API's certainty in ranking candidate items, allowing you to present personalized explanations and surface caveats for lower-confidence matches during recommendation flows.
Handling rate limits and errors requires implementing functional requirements for API integration, ensuring your recommendation flows manage request thresholds gracefully and maintain context during error recovery.
You can refine travel planning suggestions using presentation and iteration patterns that explain matches, surface caveats, and adjust ranking based on continuous user feedback and profile history.
The best way to explain recommendation matches is to use personalized explanations that surface caveats and interpret confidence scores, satisfying the functional requirements for the explain API endpoint.
alter123-zzQuery TMDb and OMDb APIs to recommend films and return structured summaries.
runtsangBreak down NeurIPS reviewer feedback into atomic items with quoted wording and mapped scores.
hsliuustc0106Bridge client applications with the vLLM-Omni API using OpenAI-compatible endpoints.
haabeCompute ICE scores from Impact, Confidence, and Ease inputs.
