seo-writer
强制先收集并锁定事实表，适合需要合规与可审计的内容团队。
基于先验数据与工具核验的可查证 SEO 文案写作，避免虚构数据与断言。
写面向搜索排名的可查证内容：文章、着陆页、产品文案和 meta 标题描述等，但严格先把事实和数据核实并锁定，写作时不凭空捏数据。当你需要上线能靠得住的 SEO 内容、避免虚假引用或伪造统计时用，流程是“先调研并锁定事实→闭卷写稿→事后审查”。最大特点是不允许 hallucination，把研究和写作分开以保证可验证性。
▸ 展开 SKILL.md 英文原文
Fact-checked SEO content writing that does NOT hallucinate. Blog articles, landing/service/city pages, product copy, meta titles + descriptions — anything meant to rank. Enforces a verify-then-write workflow: PROJECT data first, live TOOL data second, and zero invented statistics, prices, keyword metrics, studies, quotes, certifications, or URLs. Hallucination is treated as a WORKFLOW bug (research and writing in the same breath), fixed structurally: gather + LOCK facts → draft closed-book against the approved fact sheet → run a claims audit that catches anything that leaked in. UBERSUGGEST (UBBERSUGGEST MCP) is the natural T-tier source for keyword metrics (hard rule 4); per-project LAW (word counts, coverage lists, tone) lives in the project's `_seo/brief.md`, not this skill.
帮我安装这个 skill：https://raw.githubusercontent.com/dxxx/Bot-OS/main/skills/seo-writer/SKILL.mdcurl -fsSL "https://raw.githubusercontent.com/dxxx/Bot-OS/main/skills/seo-writer/SKILL.md"# SEO Writer — verified content, project data first Core principle: facts are gathered and LOCKED before drafting; drafting is closed-book against the approved fact sheet; a claims audit after drafting catches leaks. Never research and write in the same breath. > **Where the rules live:** this skill is universal. A project's specific LAW — target word counts, > the exact coverage/county list, allowed client claims, brand tone, locale — lives in that project's > **`_seo/brief.md`** (template: [`templates/brief.md`](templates/brief.md)), which Phase 0 ingests. > **Keyword metrics come from the tool** (UBBERSUGGEST MCP), never from memory (hard rule 4). ## Source hierarchy (higher tier alway