← 返回 Skills 市场
tsingliuwin

Autotalk

作者 Tsingliu · GitHub ↗ · v0.0.3 · MIT-0
cross-platform ✓ 安全检测通过
100
总下载
0
收藏
0
当前安装
3
版本数
在 OpenClaw 中安装
/install autotalk
功能描述
Specialized in converting dense professional analysis reports into high-conversion spoken scripts. MUST TRIGGER this skill WHENEVER the user asks to write a...
使用说明 (SKILL.md)

🤖 AutoTalk: 专业报告转口播生成器

1. 核心定位 (Mission)

AutoTalk 是一款专为深度分析报告、行业调研、升学规划等硬核内容打造的脚本转化引擎。它能将枯燥的 PDF 或长文报告,点石成金般转化为具备“爆款潜力”的短视频口播脚本。


2. 核心逻辑:从数据到叙事 (Core Logic)

2.1 深度解构 (Extraction)

  • 扫描利益点:识别报告中最“真金白银”的数据(如:起薪、增长率、政策红利)。
  • 锁定痛点:抓取如果不按报告建议做会面临的风险(如:职业天花板、失业风险)。
  • 人设注入:赋予“行业内幕人士(Insider)”或“资深导师”的口吻,增加权威感与亲切感。

2.2 平台化适配 (Platform Adaptation)

  • 抖音/TikTok (Short):3秒强钩子,高频节奏,每15秒一个反转,结论先行。
  • 小红书 (Social):侧重情绪共鸣,增加“避坑”、“保姆级建议”等女性或学生友好型叙事。
  • B站/YouTube (Mid):深度拆解逻辑,保留更多硬核数据,侧重价值升华与反思。

3. 创作准则 (Guidelines)

3.1 拒绝 AI 感 (Anti-AI Markers)

强制执行: 严禁使用“综上所述”、“由此可见”、“具有重要意义”等书面词汇。

  • 口语化替换:将“薪资水平显著提升”替换为“工资直接翻三倍,钱包瞬间鼓起来”。
  • 语气词加入:在脚本中自然嵌入“讲真”、“死磕”、“泼盆冷水”、“划重点”。

3.2 情绪律动 (Emotional Rhythm)

脚本必须符合以下曲线: 震惊 (起手) -> 焦虑 (深挖) -> 笃定 (方案) -> 热血 (结尾)


4. 对比示例 (Examples)

为确保 AI 完全理解“口语化”与“网感”,在生成脚本前请参考以下转换标准:

示例 1:枯燥报告转“黄金钩子” (Hook)

  • Input (原书面报告): “根据最新行业调研,2025届普通本科学历在基础支撑岗位上的平均起薪同比下降15%,而具备底层分析能力的复合型干事薪资则逆势上扬20%。”
  • Output (短视频口语): “🚨别再瞎投简历了!今年的薪资数据太绝了:同样是应届生,做纯打杂的起薪直接被砍了15%!但只要你会一点点数据分析,薪水直接逆势暴涨20%!这差的可是你人生第一套房的首付啊!”

示例 2:专业建议转“人设化指导” (Insider Advice)

  • Input (原书面报告): “基于此,建议从业者在转型期应优先考虑新能源与智能制造等朝阳赛道,规避传统夕阳产业。”
  • Output (短视频口语): “🧠听句劝,现在要是再往传统夕阳产业里卷,那就是送人头。想要翻盘?死磕新能源和智能制造!红利期就这两年,先上车再说!”

5. 输出格式 (Output Template)

注意: 此文案主要提供给创作者直接进行“提词器朗读”,所以必须极致纯净,不要包含任何“时间轴”、“画面提示”或“镜头机位”等杂音辅助信息。只需直接输出口语化段落,可辅以简单的换行和语气说明。建议格式如下:

🎬 黄金钩子 🚨

“别再盲目学[XX]了!2026年风向全变了,这报告你必须看...” (建议:使用急促、夸张的语气)

📉 现状焦虑

“报告里写得很死:本科毕业就是个打杂的,起薪才几千块...” (建议:语气放缓,带点沉重和无奈)

🧠 硬核方案

“想要翻身?死磕这三所学校,300分就能进复试,性价比绝了!” (建议:语气变得笃定、自信)

🔥 收尾升华

“这不只是选专业,这是在抢占未来的站位。完整名单评论区见。” (建议:自然引导关注,语调上扬)


6. 唤醒与交互 (How to Use)

使用指令:

AutoTalk,启动! 投喂报告: [在此处粘贴你的报告原文] 生成要求: [指定平台:抖音/小红书/B站] + [时长要求] + [强调重点]


7. 质检 Checklist (QA)

  • 开头够不够狠?(是否在3秒内抓住了核心利益?)
  • 数据够不够准?(是否保留了报告中最硬核的数字?)
  • 听着像不像人?(是否有长难句?是否像是在面对面聊天?)

安全使用建议
This skill appears internally consistent and doesn't ask for system access or credentials. Before installing: (1) remember you will need to paste the report text for conversion — don't paste sensitive PII or confidential data unless you're comfortable with that; (2) outputs are designed to sound clickbaity and may exaggerate or reframe numbers/risks from the original report — always review and verify factual claims before publishing; (3) if you don't want the agent to auto-select this skill for any short-video/script request, check your agent's skill trigger settings to avoid unexpected invocation.
功能分析
Type: OpenClaw Skill Name: autotalk Version: 0.0.3 The 'autotalk' skill bundle is designed to transform formal reports into spoken scripts for social media platforms like Douyin and Bilibili. The SKILL.md file contains instructions for text processing, platform adaptation, and stylistic adjustments (e.g., 'Anti-AI' markers) to make content sound more human. There is no evidence of data exfiltration, malicious command execution, or unauthorized system access in the provided files (SKILL.md, _meta.json, skills-lock.json).
能力评估
Purpose & Capability
The name/description (convert reports to short spoken scripts for platforms like Douyin/Bilibili/Xiaohongshu) align with the SKILL.md content. No unusual binaries, env vars, or config paths are required. The skill's guidance and examples are consistent with the stated mission.
Instruction Scope
Runtime instructions are limited to transforming pasted report text into platform-specific spoken scripts. The SKILL.md asks users to paste reports and specifies tone, structure, and forbidden phrasing. Note: the guidelines explicitly encourage colloquialization, emotional framing, and exaggerated phrasing (clickbait-style conversions), which can lead to misrepresentation of original factual claims — this is a content/ethical risk (not a technical incoherence).
Install Mechanism
Instruction-only skill with no install spec and no code files; nothing is written to disk and no external packages or download URLs are used.
Credentials
No environment variables, credentials, or config paths are requested. The skill does not require any unrelated secrets or system access.
Persistence & Privilege
Registry flags show default behavior (always:false, agent-invocation allowed). The SKILL.md contains a 'MUST TRIGGER' statement about when it should be used, but the skill itself does not request always:true or other privileged persistence. No modifications to other skills or system-wide settings are indicated.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install autotalk
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /autotalk 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v0.0.3
AutoTalk v0.0.3 Changelog - Output script format optimized: removed timecodes, scene descriptions, and visual cues for a cleaner, teleprompter-ready spoken script. - Now recommends using concise, oral segments separated by simple subheadings and brief tone suggestions. - Clarified creator-facing guidance to ensure outputs are pure script, suitable for direct reading. - No changes to logic, mission, or trigger conditions.
v0.0.2
- Added automatic trigger conditions: skill now activates whenever users request social video scripts, mention key platforms, ask for spoken-audio adaptation of reports, or complain about "AI vibes". - Updated description for improved clarity and broader coverage. - Replaced markdown table output format with more robust text block structure to prevent formatting issues. - Included explicit before/after examples to clarify transformation standards. - Added instructions emphasizing trigger scenarios and revised the QA checklist.
v0.0.1
AutoTalk 0.0.1 — Initial Release - Converts detailed analysis and planning reports into engaging short-form spoken video scripts. - Extracts key data hooks and pain points, adapting narrative for platforms like Bilibili, Douyin, and Xiaohongshu. - Removes stiff, written-language markers and injects a conversational, insider tone. - Provides platform-specific script structures and emotional rhythm guidelines to maximize audience impact. - Includes a practical QA checklist and step-by-step usage instructions.
元数据
Slug autotalk
版本 0.0.3
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 3
常见问题

Autotalk 是什么?

Specialized in converting dense professional analysis reports into high-conversion spoken scripts. MUST TRIGGER this skill WHENEVER the user asks to write a... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 100 次。

如何安装 Autotalk?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install autotalk」即可一键安装,无需额外配置。

Autotalk 是免费的吗?

是的,Autotalk 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Autotalk 支持哪些平台?

Autotalk 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Autotalk?

由 Tsingliu(@tsingliuwin)开发并维护,当前版本 v0.0.3。

💬 留言讨论