EdgeKnowledge_Skill MCP — Cas d'usage, Installer & Démo en direct

Pourquoi l'utiliser

Fonctionnalités clés

Three-tier browser routing: local Playwright, Vercel agent-browser, Google Gemini MCP
Anti-detection systems: random delays, login limits, CAPTCHA bypass
V2 Hunter Mode: value signal detection, auto-reply templates, resource downloading
Three-layer content analysis: edge knowledge, power users, resources
Structured Markdown report generation with actionable insights

Démo en direct

Aperçu en pratique

edgeknowledge-skill.replay ▶ prêt

0/0

Installer

Choisissez votre client

~/Library/Application Support/Claude/claude_desktop_config.json · Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "edgeknowledge-skill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/1596941391qq/EdgeKnowledge_Skill",
        "~/.claude/skills/EdgeKnowledge_Skill"
      ],
      "_inferred": true
    }
  }
}

Ouvrez Claude Desktop → Settings → Developer → Edit Config. Redémarrez après avoir enregistré.

~/.cursor/mcp.json · .cursor/mcp.json

{
  "mcpServers": {
    "edgeknowledge-skill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/1596941391qq/EdgeKnowledge_Skill",
        "~/.claude/skills/EdgeKnowledge_Skill"
      ],
      "_inferred": true
    }
  }
}

Cursor utilise le même schéma mcpServers que Claude Desktop. La config projet l'emporte sur la globale.

VS Code → Cline → MCP Servers → Edit

{
  "mcpServers": {
    "edgeknowledge-skill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/1596941391qq/EdgeKnowledge_Skill",
        "~/.claude/skills/EdgeKnowledge_Skill"
      ],
      "_inferred": true
    }
  }
}

Cliquez sur l'icône MCP Servers dans la barre latérale Cline, puis "Edit Configuration".

~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "edgeknowledge-skill": {
      "command": "git",
      "args": [
        "clone",
        "https://github.com/1596941391qq/EdgeKnowledge_Skill",
        "~/.claude/skills/EdgeKnowledge_Skill"
      ],
      "_inferred": true
    }
  }
}

Même format que Claude Desktop. Redémarrez Windsurf pour appliquer.

~/.continue/config.json

{
  "mcpServers": [
    {
      "name": "edgeknowledge-skill",
      "command": "git",
      "args": [
        "clone",
        "https://github.com/1596941391qq/EdgeKnowledge_Skill",
        "~/.claude/skills/EdgeKnowledge_Skill"
      ]
    }
  ]
}

Continue utilise un tableau d'objets serveur plutôt qu'une map.

~/.config/zed/settings.json

{
  "context_servers": {
    "edgeknowledge-skill": {
      "command": {
        "path": "git",
        "args": [
          "clone",
          "https://github.com/1596941391qq/EdgeKnowledge_Skill",
          "~/.claude/skills/EdgeKnowledge_Skill"
        ]
      }
    }
  }
}

Ajoutez dans context_servers. Zed recharge à chaud à la sauvegarde.

claude mcp add edgeknowledge-skill -- git clone https://github.com/1596941391qq/EdgeKnowledge_Skill ~/.claude/skills/EdgeKnowledge_Skill

Une seule ligne. Vérifiez avec claude mcp list. Supprimez avec claude mcp remove.

Cas d'usage

Usages concrets : EdgeKnowledge_Skill

How to research a niche topic across multiple forums

👤 Researchers and analysts gathering competitive intelligence ⏱ ~30 min intermediate

Quand l'utiliser : You need to understand what a community is saying about a specific topic.

Prérequis

EdgeKnowledge skill installed — chmod +x install.sh && ./install.sh
Claude Code CLI and Python 3.8+ — Install both before running the installer

Déroulement

Define your research target

Research what the indie hacker community on Reddit is saying about AI-powered SaaS tools. Focus on pricing strategies and customer acquisition.✓ Copié

→ Skill identifies relevant subreddits and begins crawling
Review the report

Generate a structured report from the crawled data. Highlight the top strategies and most-mentioned tools.✓ Copié

→ Markdown report with categorized findings and source links

Résultat : A structured research report with actionable insights from community discussions.

Pièges

Crawling too aggressively triggers rate limits — The built-in anti-detection handles this, but avoid targeting too many communities at once

Combiner avec : filesystem

Combinaisons

Associez-le à d'autres MCPs pour un effet X10

edgeknowledge-skill + filesystem

Save research reports and downloaded resources to organized folders

Research the latest trends in the SEO community and save the report to ~/research/seo-trends.md.✓ Copié

Outils

Ce que ce MCP expose

Outil	Entrées	Quand appeler	Coût
Forum crawler	target forums, search terms	Crawl forums for content matching your research criteria	0 (local) or API costs for Gemini tier
Report generator	crawled data	Generate a structured analysis report from crawled data	0

Coût et limites

Coût d'exécution

Quota d'API: Depends on routing tier — local Playwright is free, Gemini API has its own quotas
Tokens par appel: 1000–5000 tokens per research session
Monétaire: Free for local browsing; Gemini API tier may have costs
Astuce: Use the local Playwright tier for most tasks — only fall back to Gemini for CAPTCHA-heavy sites.

Sécurité

Permissions, secrets, portée

Stockage des identifiants : Forum credentials stored locally for session reuse — keep in env vars, not in skill config

Sortie de données : Connects to target forums via browser. Gemini tier routes through Google API.

Ensure your crawling complies with each forum's Terms of Service.
Anti-detection features are designed for respectful automated access, not for circumventing explicit blocks.
Downloaded resources may be subject to copyright — verify usage rights.

Dépannage

Erreurs courantes et correctifs

Browser fails to launch

Ensure Playwright browsers are installed: npx playwright install chromium. Check that Python 3.8+ is available.

Vérifier : python3 --version && npx playwright --version

CAPTCHA blocking access

The skill includes CAPTCHA bypass, but some sites may still block. Try again later or use a different routing tier.

Empty results from crawl

The target forum may require login. Provide credentials or log in manually first.

Vérifier : Try accessing the forum URL in a regular browser

Alternatives

EdgeKnowledge_Skill vs autres

Alternative	Quand l'utiliser	Compromis
fetch MCP	You just need to read a specific forum page without deep crawling	No anti-detection, no deep crawling, but simpler and faster

Plus

Ressources

📖 Lire le README officiel sur GitHub

🐙 Voir les issues ouvertes

🔍 Parcourir les 400+ serveurs MCP et Skills