LinkedIn Outreach Skill
mukulcodezz/linkedin-outreach-skill
Outreach LinkedIn personalisado e inteligente, orientado por nutrição, não por spam
Ver no GitHubSkill de agente Claude que transforma screenshots de perfil LinkedIn em mensagens de outreach personalizadas e baseadas em nutrição. Preencha um arquivo de config com informações da sua empresa e use para qualquer indústria, priorizando taxa de resposta sobre volume de mensagens.
Instalação
npx convez-ai mukulcodezz/linkedin-outreach-skill1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Gerar mensagens de alcance personalizadas a partir de perfis LinkedIn
- Automatizar cadência de follow-ups de nurture (5-7-14-21-30 dias)
- Classificar e pontuar leads por qualificação com base em sinais de atividade
- Lidar com objeções e respostas com framework estruturado
- Gerar comentários de valor para aquecer leads antes de enviar DM
Veredito

Skill bem arquitetada, focada em qualidade sobre quantidade e totalmente configurável via arquivo único. Excelente para times de vendas que querem automação inteligente respeitando termos do LinkedIn. Requer disciplina do usuário para não virar spam.
README
A tradução para o português deste README ainda está na fila.
LinkedIn Outreach Skill
🌐 Live page: https://mukulcodezz.github.io/linkedin-outreach-skill/
A Claude Agent Skill that turns LinkedIn profile screenshots into personalized, scored, nurture-first outreach — for any business. Fill one config file with your company info and it works for your industry.
It thinks like a good networker, not a spam bot. The north star is Reply Rate > Meetings > Sales > Messages Sent — so the default first message never pitches.
What it does
You share a LinkedIn profile screenshot; the skill:
- Reads the profile (headline, About, banner, Featured, recent activity).
- Classifies the prospect into a role/tier.
- Scores the lead (activity-first, 1–10, Hot/Warm/Cold) with trigger boosts (new job, hiring, funding…).
- Drafts a nurture DM (80–120 words, no pitch) — plus an optional pitch variant only for hot leads with a strong buying signal.
- Runs follow-ups on a nurture cadence (Day 0 → 5–7 → 14 → 21 → 30).
- Handles replies with an objection framework and a handoff to your closing channel.
- Generates value comments to warm leads before DMing.
- Logs every prospect and auto-prunes old entries.
Every message passes a tone checker that rejects spammy, robotic, or salesy phrasing before you ever see it.
Quick start
# 1. Clone into your Claude skills folder
git clone https://github.com/<you>/linkedin-outreach-skill.git ~/.claude/skills/linkedin-outreach
# 2. Fill in your business
# Edit config.md — replace every [BRACKETED] placeholder. See SETUP.md.
# 3. Use it in Claude Code
# Paste a LinkedIn profile screenshot and say: "draft an outreach DM"
# (or invoke the skill by name: linkedin-outreach)
Windows path: C:\Users\<you>\.claude\skills\linkedin-outreach.
The one file you edit
config.md holds 100% of your business specifics — company, offerings→outcomes, proof points, ICP, role tiers, buying triggers, voice, and handoff channel. Everything in references/ is generic methodology that reads from it. See SETUP.md.
Structure
linkedin-outreach/
├── SKILL.md # the engine: modes + flow (don't need to edit)
├── config.md # ← FILL THIS (your whole business)
├── SETUP.md # fill-in guide
├── references/
│ ├── scoring.md # lead scoring rubric + tiers
│ ├── triggers.md # buying triggers → score boost + hooks
│ ├── followups.md # nurture follow-up cadence
│ ├── reply_handling.md # objection bank + handoff
│ ├── comments.md # value-comment generator
│ ├── tone_checker.md # pre-send quality gate
│ ├── constraints.md # length/format rules
│ └── examples.md # how-to + a filled fictional example
└── logs/
└── sent_log.md # prospect tracker (auto-pruned)
Modes
| You share / ask | Mode |
|---|---|
| A profile screenshot | New DM |
| "Follow up with [name]" | Follow-Up |
| A screenshot of their reply | Reply Handling |
| "Comment on this post" + screenshot | Comment |
Notes
- Designed for outreach to existing 1st-degree connections (LinkedIn restricts DMs to non-connections).
- Use it responsibly: personalized, genuine, low-volume outreach — not mass automation. Respect LinkedIn's terms.
config.mdandlogs/sent_log.mdwill contain your private data once filled. If you fork publicly, keep them out of commits (see.gitignore).
License
MIT — see LICENSE.