Skip to content

feat: add recsys-pipeline-architect skill#1163

Open
mturac wants to merge 1 commit into
vercel-labs:mainfrom
mturac:add-recsys-pipeline-architect
Open

feat: add recsys-pipeline-architect skill#1163
mturac wants to merge 1 commit into
vercel-labs:mainfrom
mturac:add-recsys-pipeline-architect

Conversation

@mturac
Copy link
Copy Markdown

@mturac mturac commented May 16, 2026

Adds a new skill: recsys-pipeline-architect.

A vendor-agnostic Agent Skill (agentskills.io standard) for designing composable recommendation, ranking, and feed pipelines using the six-stage Source → Hydrator → Filter → Scorer → Selector → SideEffect framework popularized by xAI's open-sourced For You algorithm (Apache 2.0; this skill is MIT and is an independent reimplementation of the pattern).

Activation

Triggers on any "top K items for (user, context)" problem: content feeds, search ranking, RAG retrieval rerankers, task prioritizers, notification triage, ad selection. The skill walks the agent through eight clarifying steps and emits a runnable scaffold in the user's stack.

Why a vendor-agnostic skill belongs here

Like the existing find-skills skill, this one is platform-neutral — the SKILL.md is consumed identically by Claude Code, Codex, Cursor, Gemini CLI, OpenCode, Windsurf, and the other agents that npx skills add supports.

Upstream

This is a single-file adapter for the upstream repository: https://github.com/mturac/recsys-pipeline-architect (v0.1.0, MIT). Upstream ships 5 load-on-demand reference docs (interfaces in 4 languages, multi-action scoring, candidate isolation, filter cookbook with 12 patterns, scorer cookbook) and 3 runnable example scaffolds (Strapi v5/TS/Jest, Go with generics, Python/FastAPI/pytest — every one green: 9/9 tests passing total).

Cross-platform install via skills.sh: npx skills add mturac/recsys-pipeline-architect.

Adds a skill for designing composable recommendation, ranking, and feed
pipelines using the six-stage Source→Hydrator→Filter→Scorer→Selector→
SideEffect framework popularized by xAI open-sourced X For You algorithm.
Independent MIT reimplementation; surfaces multi-action vs single-score,
candidate isolation vs joint, and online vs offline batch trade-offs
explicitly.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant