菜单
Detect winners, decaying posts, striking-distance keywords, and surprise wins. Computes week-over-week deltas from GSC + Fathom + Kit snapshots across the current and previous weeks.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/kkoppenhaver/weekly-marketing-report --skill analyze-content-performance命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | analyze-content-performance |
| description | Detect winners, decaying posts, striking-distance keywords, and surprise wins. Computes week-over-week deltas from GSC + Fathom + Kit snapshots across the current and previous weeks. |
Reads the current and previous week's search-console.json, fathom.json, kit.json, and posts.json. Emits insights.content_perf.
python .claude/skills/analyze-content-performance/scripts/analyze.py [--week 2026-W18]
Requires at least two consecutive weeks of fetcher data for WoW analysis. First-run output is partial — note this in the report.
See plan.md §3.8 for the full schema. Sections:
headline — top-line WoW deltas (clicks, signups, mover count)winners_this_week — top movers updecaying_posts — posts trending down 4+ weeksstriking_distance_keywords — keywords ranking position 8–20 with ≥50 weekly impressionssurprise_winners — posts ranking well for keywords they didn't declaresignup_attribution — signups by source page (when Fathom/Kit data permits)impressions × (1 / position).