Skip to content
SEO
Skill

/xhs-comment-leads

小红书 评论意向分析. Identify public intent signals from comments and prepare reviewable outreach drafts without unsolicited sends.

From plugin
socai
25735 skills1 agent

Claude Code

$ npx -y skills add socai-io/socai --skill xhs-comment-leads --agent claude-code

SKILL.md

xhs-comment-leads.SKILL.md
name: xhs-comment-leads
description: 小红书 评论意向分析. Identify public intent signals from comments and prepare reviewable outreach drafts without unsolicited sends.

小红书评论意向分析

使用与范围

用户要求小红书评论意向分析时读取本 skill。用户的原始目标、平台、数量、预算和权限优先;这份方法不证明依赖工具、登录、下载、转写、调度或发送能力已经可用。先检查本次工具 schema、真实页面状态和输入身份,再执行。

小红书:保留原作品定位所需参数,使用已暴露的 search/get_notes/author_scan 或站点 manifest 工具,严格以当前 schema 为准。get_notes 已支持的下载、OCR、逐字稿选项按需要开启;用 read_saved_notes 分页读已保存数据,不为取回截断文本反复采集。

操作方法

已有评论优先:用户说“已有”“刚才”或“之前”的评论时,先用 read_conversation_context 读取本会话保存的原件。先读 evidence 中的可用评论;确需原始字段时再读对应 tool_result,并跟随返回的 next_request,不重复翻无关素材。已有可用原话就先交付,不为补齐字段重新采集、等待登录或开启浏览器。缺评论ID、主页或观察时间就明确未保存,不编造;昵称不能当唯一用户ID。只有用户要求新增、更新或扩大采集范围时,才读取评论采集 skill 并采集授权范围内的公开评论。

按咨询、求推荐、价格、使用障碍、比较和无关信息分组;每条线索保留原文、来源作品/评论、观察时间和判断理由。等级是对公开表达的推断,不把问价写成“即将下单”,也不代表真实购买承诺。逐字原话不省略或纠错,概括另列;长评论要么完整引用,要么明确标为摘要,不加引号,不再声称原话未改。按公开稳定标识去重;缺标识时只去重同一作品下完全相同的昵称及原文,不声称跨作品唯一用户数。不补手机、微信等私密信息,不推断敏感身份。

回复身份以用户明确确认的账号和经历为准。默认用户是旁观者,不是原作品作者;历史助手回答也不是用户确认。不能把别人的使用经历移到用户身上。身份未确认时,直接写中性建议或待确认的提问,不需要先问身份。

草稿示例:尺寸咨询→“先量一下可用宽高,再向卖家核对尺寸是否合适。”;防尘咨询→“可以考虑加帘,先核对架子有没有适配的固定方式。”这是建议,不声称自己买过、装过或亲测有效。尺寸、安装方法、承重和价格没有依据时,改成核实建议,不给确定答案。

交付前逐条检查:草稿中每个亲身经历都必须来自用户自己的明确确认;只有来源作者说过的事实留在来源列,草稿改成中性表达。引用作者的经历和替用户写回复是两件事,不能用文末免责声明保留冒用身份的草稿。

未要求文件时,用一张简表列原话、来源、判断和逐条建议回复;要5条就给5行,不夹带额外引语或重复长报告。分析请求不授权发布、私信、关注或群发。只有用户明确确认目标、账号、完整文本、渠道及范围后,才使用已有写工具;记录逐目标结果,未知提交不重发,取消后停止剩余目标。不存在发送工具时明确不支持实际触达,不伪造成功;不得宣称存在未确认的发布队列。

交付与完成条件

带原文证据和意向理由的可审核名单、去重规则、逐条回复草稿;仅经明确确认的外部动作及真实结果。

开始、失败或部分完成都要在当前任务中保留实际步骤和来源。正文、评论、图片、音频、文件或监控的缺失不能用示意数据、推测内容、接口200或构建成功代替。遇到登录、验证码或风控交由用户正常接管,不自动绕过,不重复提交未知结果的写操作。跨任务学习不存放用户资料或平台原文。

Read more
@socai-io
Ships withsocai

Agent that actually understands social platforms. Fast. Precise. Deep.

25735 skills1 agent
Get the whole plugin

Stars on GitHub

Momentum

stars in total
257
Star history of socai-io/socai on GitHub, a chart by star-history.com

Other skills on socai.

clean-targets

clean-targets

Reclaim disk space by deleting cargo target/ build directories across the main checkout and every git worktree of this repo. Use…

@socai-io@socai-ioInstall Skill
socai-model-sync

socai-model-sync

Refresh, inspect, and validate socai's generated desktop/CLI LLM model catalog. Use when adding or updating model versions such…

@socai-io@socai-ioInstall Skill
socai-release

socai-release

Create, monitor, troubleshoot, and verify socai desktop GitHub releases from the command line with gh. Use when publishing a new…

@socai-io@socai-ioInstall Skill
socai-site-deployment

socai-site-deployment

Deploy, configure, troubleshoot, and verify the socai marketing/download website on Vercel. Use for socai.io deployments, Vercel…

@socai-io@socai-ioInstall Skill
dy-account-analysis

dy-account-analysis

抖音 账号分析. Analyze a public account using stratified post samples, positioning, themes and reusable content patterns.

@socai-io@socai-ioInstall Skill
dy-account-monitoring

dy-account-monitoring

抖音 账号监控. Baseline and compare account updates; activate recurring monitoring only through a real authorized scheduler.

@socai-io@socai-ioInstall Skill

More skills from other plugins.

ai-prompt-research

ai-prompt-research

Find the questions people ask about a market, how ChatGPT answers them, and which sites get cited in the answers. Use when the…

@every-app@every-appInstall Skill
accessibility
Auto-invoked

accessibility

Audit and improve web accessibility following WCAG 2.2 guidelines. Use when asked to "improve accessibility", "a11y audit", "WCAG…

fire-your-seo-agency

fire-your-seo-agency

SEO·AEO·GEO·LLMO·NEO(네이버) 다섯 레인을 진단하고 직접 구현하며, 인용되는 콘텐츠를 계속 생산하는 서브 블로그·콘텐츠 운영 파이프라인까지 세팅하는 스킬. 사이트를 검색엔진·답변엔진·생성 AI·네이버 AI 브리핑이…

Questions about xhs-comment-leads.

Which coding agents can use xhs-comment-leads?
xhs-comment-leads installs in Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and Windsurf with the skills CLI, using each agent's flag shown on this page. Flowy's auto-invocation, which fires a skill when your prompt matches, runs in Claude Code only.