在 Manus 中运行任何 Skill
一键导入
一键导入
一键在 Manus 中运行任何 Skill
开始使用$pwd:
content-creation
// Mastering High-Quality Thai Content Generation with Adaptive RAG v3 Evidence
$ git log --oneline --stat
stars:1
forks:0
updated:2026年3月24日 22:06
SKILL.md
// Mastering High-Quality Thai Content Generation with Adaptive RAG v3 Evidence
[HINT] 下载包含 SKILL.md 和所有相关文件的完整技能目录
| name | content-creation |
| description | Mastering High-Quality Thai Content Generation with Adaptive RAG v3 Evidence |
This skill defines the technical benchmarks and linguistic rules for generating "Masterpiece" Thai content using the Adaptive RAG v3 architecture.
Content must be Evidence-First, Human-Centric, and Contextually Grounded. Never generate fluff. Every claim must have an underlying trace to the research data.
Before writing, always ensure the research data is cleaned and synthesized:
agentFilterFeed to select high-signal posts.buildSearchPlan to ensure both Web (Tavily) and X data are present.Fact Sheet with Verified Facts, Open Questions, and Community Signals.Never jump straight to drafting. Use buildContentBrief to define:
| Rule | Bad AI-Thai | Good Native Thai | Rationale |
|---|---|---|---|
| Space | "บริษัทAppleเปิดตัวสินค้า!" | "บริษัท Apple เปิดตัว..." | Improved readability. |
| Passive | "มันถูกมองว่าเป็น..." | "หลายคนมองว่า..." | Professional tone. |
| Hype | "โลกจะไม่เหมือนเดิมอีกต่อไป" | "แคมเปญนี้ได้เปลี่ยนภาพตลาด..." | Professional credibility. |
| Emoji | "🚀🔥⚡️" | "🔥" (Max 3-4 per post) | Avoid spammy look. |
GrokService.js: generateStructuredContentV2 (Primary entry point)TwitterService.js: Evidence curation logic