ray

Routes user tasks to specialized content production, consulting, and orchestration skills.

160|18|Updated Jul 13, 2026
One-click install
npx skills add https://github.com/imraywang/rayskills --skill ray-imraywang
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ray
Source: https://github.com/imraywang/rayskills/tree/main/skills/ray
Command: npx skills add https://github.com/imraywang/rayskills --skill ray-imraywang

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? When you have a real task but don't know which specialized skill to use, this router reads your context and picks the single most appropriate next step, or runs a confirmed end-to-end content pipeline from idea to verified platform drafts. ## Core Features & Use Cases - Intent Routing: Reads your current situation and dispatches to the right member skill covering Obsidian knowledge bases, long-form writing, translation, covers, WeChat drafts, X Articles, oral video scripts, B-roll, consulting, and multi-model collaboration. - Content Pipeline Orchestration: When you explicitly request end-to-end production, it chains ray-writer, ray-cover, ray-wechat, and ray-x-article with stage gates, stopping at verified drafts and never publishing. - Authorization Grading: Classifies every platform action as local_only, draft_write, or public release based on your exact words, so vague phrases like "push to platforms" never become unauthorized writes or publications. - Use Case: You say "turn this idea into an article, make covers, and save it to my X Articles draft box" — the router runs the full pipeline and stops at a verified draft without publishing. ## Quick Start Type /ray followed by a description of your real task or situation, such as "/ray turn this idea into a WeChat article draft", and let it choose and execute the right step.

Frequently Asked Questions about ray

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I route a task to the right AI skill without knowing skill names?▼

Invoke /ray followed by a plain description of your situation or goal. The router reads your context, matches intent against its routing table, and selects the single most appropriate skill, explaining why it was chosen.

How do I run an end-to-end content pipeline from idea to platform drafts?▼

Explicitly state the full goal, such as writing an article, generating covers, and saving to X Articles or WeChat draft boxes. The router then chains ray-writer, ray-cover, and the platform skills with stage gates, stopping at verified drafts.

Does the router ever publish content to WeChat or X automatically?▼

No. Public publishing is explicitly outside the pipeline. The authorization ladder only permits local artifacts or verified drafts, and vague phrases like "push to platforms" are downgraded to local-only output with a confirmation question.

What happens when my request is too vague to route?▼

The router does not guess or force a dispatch. It asks exactly one clarifying question targeting the minimum information needed to identify the right skill, then proceeds once you answer.

Can draft authorization for one platform carry over to another?▼

No. Authorization is judged per platform from your current message. Approval to write a WeChat draft never grants X draft access, and past permissions, login state, or fixed whitelists never expand the current authorization.