Token导航 LogoToken导航TokenDH.com
开发操作浏览器github未标认证来源可访问许可证需确认审计提醒

high-fidelity-extraction高保真提取

Agent Skill

high-fidelity-extraction 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

539

周安装

22

GitHub Stars

17

下载量

172
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:high-fidelity-extraction(高保真提取)
来源仓库:https://github.com/jstarfilms/vibecode-protocol-suite
仓库路径:skills/high-fidelity-extraction
安装命令:
npx skills add https://github.com/jstarfilms/vibecode-protocol-suite --skill high-fidelity-extraction
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/jstarfilms/vibecode-protocol-suite --skill high-fidelity-extraction

简介

用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息。

  • 适合围绕仓库状态、代码变更或协作事项进行整理。
  • 可结合来源仓库和原始 README 核验具体用法。
  • 安装方式:npx skills add https://github.com/jstarfilms/vibecode-protocol-suite --skill high-fidelity-extraction。
  • 建议确认权限范围、维护状态及是否触发联网或文件读写。

SKILL.md

High Fidelity Data Extraction Protocol

This skill enables the agent to extract deep intelligence from social media platforms (Instagram, TikTok, YouTube, etc.) and complex web environments with high precision.

🚀 Core Philosophy: The Extraction Spectrum

When tasked with extraction, always check the prompt for specific limitations. If no limitations are provided, assume Standard Level.

Extraction Capability Matrix

FeatureBasic (Standard)Deep (Advanced)Elite (Full Intelligence)
CaptionsFull Text + Hashtags+ Edited timestamps+ OCR from video overlays
CommentsTop 3-5 (Top Level)Top 10 + Threaded repliesFull sentiment & pain point mapping
EngagementLikes / Views countEngagement Rate %Share/Save estimates & Viral Velocity
Brand Intel@Mentions in caption+ Link-in-bio analysis+ Competitor comparison vs Meta Ad Library
VisualsProfile Screen/GridIndividual Post ScreenshotsUI/UX Reverse Engineering of Funnel
TechnicalURL collectionPrecise ISO TimestampsAPI Pattern Mapping & DB Schema generation

📊 Tabular Output Format

Unless the user specifies otherwise, all extracted data should be compiled into a Markdown table for maximum scannability.

Standard Template:

| Post Type | Caption Snippet | Top Comments (Synthesized) | Key Metrics | Brand/Tags |
| :--- | :--- | :--- | :--- | :--- |
| Pinned | "How to be..." | Users asking about discipline tips | 30k Likes | @Adidas, #NYC |
| Recent 1 | "Day in life..." | High praise for work/life balance | 15k Views | @CeraVe |

🧠 Smart Filtering Strategy (Instagram Reels)

When the goal involves "high-performing" content or maximizing engagement:

  1. Reels First: Navigate to the /reels/ tab immediately. View counts are not visible on the main grid but are overlayed on Reel thumbnails.
  2. Baseline Calculation:

- Extract view counts from the first 6-12 visible Reels. - Calculate the average (mean) view count.

  1. Filtration:

- Only click/extract posts that exceed this average. - This ensures we focus valuable browser resources on proven content.

🛠️ Execution Instructions for the Subagent

  1. DOM First: Never navigate blindly. Use browser_get_dom to identify the obfuscated class names for captions and comments.
  2. Surgical Navigation: To avoid 429 errors (Too Many Requests), navigate directly to post URLs once the links are gathered from the grid, rather than infinite scrolling.
  3. Expansion Logic: Always look for ... more or "View all comments" buttons and trigger them via JavaScript to ensure data completeness.
  4. Verification: Always capture a final screenshot of the "Main Extraction Target" to verify text accuracy against the visual truth.

🛑 Limitations & Compliance

  • Obey Prompt Bounds: If the user says "Only get usernames," do NOT extract captions.
  • Privacy: Respect platform TOS by focusing on public-facing data and ignoring private user details.
  • Anti-Spam: Filter out "Promote on..." or generic bot comments automatically.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

38.67%
按下载量换算67

Claude

28.85%
按下载量换算50

Cursor

17.1%
按下载量换算29

Gemini CLI

8.55%
按下载量换算15

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills