Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →为单篇或多平台自媒体创作澄清需求并生成可确认的创作简报。用于用户只给一句话选题、一个链接、一份素材、一个产品或模糊创作要求，需要确认目标受众、内容目标、核心判断、角度、深度、调性、证据、禁区和平台差异时。纯数据分析、已确认方向后的改稿和细节微调不使用此 Skill。
.claude/skills/yanhua1010-self-media-content-brief/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 9% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 5% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 2% | 0% |
按影响大小检查：
| 场景 | 优先问题 | |---|---| | 公众号 | 目标读者、内容目标、深度、核心角度 | | 小红书 | 受众身份、搜索问题、图文或视频、真实证据 | | X | 想验证的判断、受众、短帖或线程、唯一行动 | | 视频号或抖音 | 时长、出镜方式、钩子方向、可用画面 | | 多平台 | 主平台、各平台承担的任务、可投入时间 |
缺少信息但可以从素材、项目文件或公开来源得到时，先查找，不反问用户。
使用以下结构：
markdown## 创作简报 - 账号定位： - 目标受众： - 内容目标： - 主平台： - 扩展平台： - 内容形态： - 一句话核心判断： - 准备使用的证据： - 读者收益： - 深度与调性： - 明确不写： - 各平台角度： - 截止时间：
列出使用的默认值。给出简报后停下，等待用户确认或修改。方向未确认前不要写完整正文。
用户确认后：
self-media-trend-radar。self-media-platform-copywriting。self-media-short-video。| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 14,542 | 12,579 | -13% | 1 | 1 | 0% | 2,082 | 2,276 | +9% | 0 | 0 | — |
case-02 | fail→pass | 15,728 | 12,785 | -19% | 1 | 1 | 0% | 2,232 | 2,353 | +5% | 0 | 0 | — |
case-03 | fail→pass | 15,511 | 13,445 | -13% | 1 | 1 | 0% | 2,295 | 2,564 | +12% | 0 | 0 | — |
case-04 | pass→pass | 18,180 | 16,117 | -11% | 1 | 1 | 0% | 2,458 | 2,785 | +13% | 0 | 0 | — |
case-14 | pass→pass | 12,762 | 3,624 | -72% | 1 | 1 | 0% | 1,788 | 1,031 | -42% | 0 | 0 | — |
case-05 | pass→fail | 25,390 | 12,650 | -50% | 1 | 1 | 0% | 3,461 | 2,365 | -32% | 0 | 0 | — |
case-06 | pass→pass | 18,807 | 17,581 | -7% | 1 | 1 | 0% | 2,770 | 3,126 | +13% | 0 | 0 | — |
case-07 | fail→pass | 34,652 | 13,122 | -62% | 1 | 1 | 0% | 4,394 | 2,439 | -44% | 0 | 0 | — |
case-08 | fail→pass | 17,742 | 14,539 | -18% | 1 | 1 | 0% | 2,465 | 2,523 | +2% | 0 | 0 | — |
case-09 | pass→pass | 19,264 | 10,234 | -47% | 1 | 1 | 0% | 2,769 | 2,020 | -27% | 0 | 0 | — |
case-10 | fail→pass | 20,129 | 12,234 | -39% | 1 | 1 | 0% | 2,828 | 2,307 | -18% | 0 | 0 | — |
case-11 | fail→pass | 13,762 | 11,563 | -16% | 1 | 1 | 0% | 1,960 | 2,175 | +11% | 0 | 0 | — |
case-12 | fail→pass | 21,068 | 17,283 | -18% | 1 | 1 | 0% | 2,962 | 2,879 | -3% | 0 | 0 | — |
case-13 | fail→pass | 13,378 | 12,541 | -6% | 1 | 1 | 0% | 1,757 | 2,229 | +27% | 0 | 0 | — |
case-15 | pass→pass | 10,028 | 4,527 | -55% | 1 | 1 | 0% | 1,416 | 1,166 | -18% | 0 | 0 | — |
case-16 | pass→pass | 10,268 | 3,794 | -63% | 1 | 1 | 0% | 1,391 | 1,073 | -23% | 0 | 0 | — |
case-17 | pass→pass | 15,310 | 12,355 | -19% | 1 | 1 | 0% | 2,073 | 2,265 | +9% | 0 | 0 | — |
case-18 | fail→fail | 18,479 | 20,968 | +13% | 1 | 1 | 0% | 2,443 | 3,382 | +38% | 0 | 0 | — |
case-19 | fail→pass | 19,584 | 11,320 | -42% | 1 | 1 | 0% | 2,592 | 2,083 | -20% | 0 | 0 | — |
case-20 | fail→fail | 19,261 | 12,667 | -34% | 1 | 1 | 0% | 2,660 | 2,295 | -14% | 0 | 0 | — |
case-21 | fail→pass | 24,169 | 13,327 | -45% | 1 | 1 | 0% | 3,277 | 2,317 | -29% | 0 | 0 | — |
case-22 | pass→pass | 14,302 | 11,022 | -23% | 1 | 1 | 0% | 2,001 | 2,083 | +4% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +45 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.