Recommendation Pipeline Architect
Designs and scaffolds ranked feed and top-K recommendation pipelines using a six-stage pattern
- Repository
- wshobson/agents
- Stars
- 42,095
- Licence
- MIT
- Files
- 1 file
What it does
When installed, the agent helps you design systems that pick the best K items for a user or context, such as a content feed, search reranker, notification triage, or task prioritizer. It walks you through a fixed structure: gather candidates, enrich them with data, filter out ineligible ones, score the rest, select the top K, then run background tasks like logging. It can then generate working starter code in your stack, and it raises key design choices (single vs multi-signal scoring, real-time vs precomputed) for you to decide.
Great for
- Engineers building a personalized feed or ranked list in a web or data product
- Teams adding a reranking step to a retrieval-augmented generation (RAG) search system
- Developers who want a clean, testable structure for notification or task prioritization
Skip it if
- Simple sorting or filtering that does not need personalization, since the framework adds unnecessary structure
- Non-technical users wanting a finished recommender product; the output is a design and scaffold that still needs engineering and tuning
Try asking
- “Design a personalized article feed for my CMS with 50k posts, using multi-action scoring and author diversity”
- “Build a reranker that takes the top 50 chunks from my vector database and returns the best 5 for my RAG app”
- “Create a notification triage pipeline that ranks alerts per user and logs impressions without slowing responses”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
recsys-pipeline-architect.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip recsys-pipeline-architect.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip recsys-pipeline-architect.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Give the agent your items, input context, language and runtime up front so it can skip extra questions.
- Ask for the trade-offs explicitly; the skill is designed to surface product decisions rather than choose silently.
- Run any benchmarks yourself: the skill is instructed not to invent performance numbers.
What's inside
- SKILL.md
Redistributed unmodified from wshobson/agents under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.