Install any skill in seconds. Free to start, no credit card required.
Get Started Free →搜索社区、社媒、新闻热点、百科趋势、开发者生态和技术讨论时使用。
.claude/skills/opensensenova-sn-search-social-media/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-16 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 51% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 88% | 0% |
API key、token 与 cookie 统一建议写在仓库根目录 .env(参考 .env.example),并由 runtime 或用户在执行前加载为同名环境变量。脚本仍只从环境变量或显式 CLI 参数读取凭证;不要把真实密钥写入 skill payload、报告、日志或提交。
本技能给 agent 用于抓取公开热点信号。只使用免费、免注册、免 API key、非加密货币的数据源;不要要求用户提供 token、cookie、API key,也不要使用商业收费接口。
items[].url 或 browser-use 打开的页面链接作为证据。首次运行或脚本提示缺库时,使用本技能的依赖清单安装到当前 Python 环境:
bashpython3 -m pip install -r requirements.txt
不要在脚本内部自动安装依赖。若安装失败、网络不可用或包不可用,停止使用对应脚本并改用 browser-use/公开网页搜索,说明缺少依赖。
| 场景 | API 脚本 | browser-use 入口 | 说明 | | --- | --- | --- | --- | | 百科页面关注度、公众注意力辅助指标 | scripts/wikimedia_pageviews.py | https://doc.wikimedia.org/generated-data-platform/aqs/analytics-api/ | Wikimedia Pageviews 需要 User-Agent | | 开源项目热度、工程活动、开发者生态 | scripts/github_public_search.py | https://github.com/search | 只搜公共仓库和 Issue;不搜 code | | 技术问题热度、开发者需求 | scripts/stackexchange_search.py | https://stackexchange.com/search | 默认 Stack Overflow,可指定 --site | | 创业、科技社区讨论热榜 | scripts/hackernews_hotspots.py | https://news.ycombinator.com/ | 官方 API 不支持服务端搜索,脚本本地过滤 | | 指数趋势、搜索热度、地域、人群画像、需求变化 | 无 | https://index.baidu.com/ | 百度指数;网页可用,只用免费可见内容 | | 社媒舆情、热点、品牌声量、事件监测 | 无 | https://s.weibo.com/top/summary | 微博热搜/微指数;网页可用,只用免费可见内容 |
以下 scripts/... 路径相对本 skill 目录;若当前工作目录不同,先解析为绝对路径:
bashpython3 scripts/wikimedia_pageviews.py article "Artificial intelligence" --start 2026-05-01 --end 2026-05-17 python3 scripts/wikimedia_pageviews.py top --date 2026-05-17 --project en.wikipedia.org --limit 20 python3 scripts/github_public_search.py "llm agents" --type repositories --sort stars --limit 10 python3 scripts/github_public_search.py "agent framework bug" --type issues --sort comments --limit 10 python3 scripts/stackexchange_search.py "python asyncio" --site stackoverflow --sort votes --limit 10 python3 scripts/hackernews_hotspots.py --kind top --limit 10 python3 scripts/hackernews_hotspots.py --kind best --query "AI agents" --scan 150 --limit 10
wikimedia_pageviews.py 默认使用带项目 URL 的 User-Agent;如运行环境需要自定义,可设置 WIKIMEDIA_USER_AGENT,这不是 API key。
article;找当天广泛关注对象跑 top,也可用 browser-use 打开百科页面核对实体含义。--scan,不要改用非官方或收费搜索 API。脚本输出:
json{ "success": true, "query": "...", "provider": "wikimedia-pageviews|github-public|stackexchange|hackernews", "items": [ {"title": "...", "url": "...", "snippet": "..."} ], "error": null }
当 success=false 且错误提示为加密货币限制时,停止处理该请求。
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 6,685 | 8,526 | +28% | 1 | 1 | 0% | 508 | 1,807 | +256% | 0 | 0 | — |
case-16 | fail→pass | 11,461 | 3,747 | -67% | 1 | 1 | 0% | 1,581 | 2,124 | +34% | 0 | 0 | — |
case-02 | fail→fail | 6,891 | 33,441 | +385% | 1 | 1 | 0% | 389 | 4,391 | +1029% | 0 | 0 | — |
case-03 | fail→pass | 13,455 | 2,766 | -79% | 1 | 1 | 0% | 2,120 | 1,782 | -16% | 0 | 0 | — |
case-04 | fail→pass | 7,729 | 2,580 | -67% | 1 | 1 | 0% | 1,238 | 1,875 | +51% | 0 | 0 | — |
case-05 | fail→pass | 10,673 | 5,148 | -52% | 1 | 1 | 0% | 2,323 | 2,181 | -6% | 0 | 0 | — |
case-06 | fail→pass | 4,992 | 2,621 | -47% | 1 | 1 | 0% | 1,003 | 1,881 | +88% | 0 | 0 | — |
case-07 | fail→pass | 6,054 | 1,579 | -74% | 1 | 1 | 0% | 1,231 | 1,664 | +35% | 0 | 0 | — |
case-08 | fail→pass | 8,863 | 4,000 | -55% | 1 | 1 | 0% | 1,868 | 2,142 | +15% | 0 | 0 | — |
case-09 | fail→pass | 7,924 | 2,102 | -73% | 1 | 1 | 0% | 1,382 | 1,703 | +23% | 0 | 0 | — |
case-10 | fail→pass | 12,714 | 1,852 | -85% | 1 | 1 | 0% | 2,616 | 1,717 | -34% | 0 | 0 | — |
case-11 | fail→pass | 14,232 | 3,480 | -76% | 1 | 1 | 0% | 2,251 | 1,952 | -13% | 0 | 0 | — |
case-12 | fail→pass | 13,267 | 4,742 | -64% | 1 | 1 | 0% | 2,314 | 2,128 | -8% | 0 | 0 | — |
case-13 | pass→pass | 10,683 | 3,128 | -71% | 1 | 1 | 0% | 1,859 | 1,930 | +4% | 0 | 0 | — |
case-14 | pass→pass | 8,030 | 4,093 | -49% | 1 | 1 | 0% | 1,407 | 1,907 | +36% | 0 | 0 | — |
case-15 | pass→pass | 13,664 | 9,303 | -32% | 1 | 1 | 0% | 2,018 | 2,859 | +42% | 0 | 0 | — |
case-17 | pass→pass | 7,327 | 2,549 | -65% | 1 | 1 | 0% | 1,185 | 1,856 | +57% | 0 | 0 | — |
case-18 | fail→pass | 5,319 | 5,284 | -1% | 1 | 1 | 0% | 1,121 | 2,261 | +102% | 0 | 0 | — |
case-19 | fail→pass | 9,546 | 5,339 | -44% | 1 | 1 | 0% | 1,486 | 2,181 | +47% | 0 | 0 | — |
case-20 | fail→pass | 11,802 | 3,822 | -68% | 1 | 1 | 0% | 1,953 | 1,877 | -4% | 0 | 0 | — |
case-21 | fail→pass | 12,935 | 6,002 | -54% | 1 | 1 | 0% | 2,361 | 2,560 | +8% | 0 | 0 | — |
case-22 | fail→pass | 13,419 | 6,833 | -49% | 1 | 1 | 0% | 2,157 | 2,277 | +6% | 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, and 20 counted toward the lift figure. The other 2 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +73 percentage points is the difference between those two pass rates over the 20 comparable cases.
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.