Cold outreach
Marketing Intelligence pack for Claude Code - website audits, meeting prep, cold outreach, ad creative analysis
npx -y skills add veyralabsgroup/marketing-suite --skill cold-outreachAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Research a prospect and generate a personalized cold outreach sequence. First message, two follow-ups, and a break-up message. Based on real signals from their website, not generic templates.
SKILL.md
5.8 KB, ~1.4k tokens by cl100k_base, as published. Nobody here has run it
Cold Outreach
Research a prospect company and generate a 4-message outreach sequence personalized to their specific situation. Not templates - actual personalization based on what you find.
What this does
Three phases:
- Prospect Research - Extract signals from their website, marketing activity, and context
- Pain Signal Identification - Find the specific angle that will resonate
- Sequence Generation - Write 4 messages: opener, follow-up 1, follow-up 2, break-up
How to use
Pass a company URL or LinkedIn URL:
/cold-outreach https://empresa.com
Or with context:
/cold-outreach https://empresa.com - contacto: Ana Lopez, directora de marketing, LinkedIn: linkedin.com/in/analopez
Phase 1 - Prospect Research
Collect the signals that make outreach feel personal:
From their website:
- What they do and who they serve
- What they are clearly investing in (design, content, ads)
- What is clearly missing or underperforming (compared to what you'd expect)
- Any recent changes (new product, rebrand, expansion)
- Tech stack signals (no pixel = cannot run ads; no analytics = flying blind)
From search:
- Recent news, press, blog posts
- Job postings (reveals priorities and pain)
- Social media activity (are they active? what topics?)
- Reviews or testimonials (reveals what customers value)
From LinkedIn (if profile provided):
- Contact's role and responsibilities
- Recent activity / posts
- Time in current role (new role = often looking to prove impact)
- Shared connections or context
Phase 2 - Pain Signal Identification
Pick ONE primary pain angle. Not five angles - one.
The best angle is:
- Specific to them (you saw something on their site / LinkedIn)
- Connected to a business outcome they care about (revenue, leads, time)
- Something the agency can actually help with
Angle selection:
| Finding | Angle |
|---|---|
| No conversion tracking | "You're spending on [channel] with no way to measure return" |
| No Meta Pixel | "Your audience data is building in competitors' accounts, not yours" |
| Inactive social (3+ months) | "Your social presence is sending the wrong signal to new visitors" |
| No case studies | "Prospects who visit your site can't see proof - that's a closing problem" |
| Old website | "Your website is costing you credibility with every new visitor" |
| SEO competitors ranking above them | "X and Y are capturing searches you should be winning" |
| Hiring for marketing | "You're scaling marketing - this is the moment to get the foundation right" |
What to avoid:
- "I loved your website" (nobody believes it)
- "I noticed you could improve your SEO" (too generic)
- "We work with companies like yours" (irrelevant)
- "I wanted to reach out because..." (filler)
Phase 3 - Sequence Generation
Generate 4 messages. Each has a different purpose.
Message 1 - The Opener
Purpose: Get a reply. One ask. One specific observation.
Format:
Subject: [specific, not clickbait - max 7 words]
[Their name],
[One observation from research - something specific to them]
[One sentence connecting it to a business outcome]
[One question or one offer - not both]
[Name]
Rules:
- Under 80 words
- No attachment, no portfolio link in first message
- The observation must be specific enough that they think "they actually looked at my site"
- One ask only: either a question OR a link/offer, never both
Example:
Subject: Your Meta Pixel isn't firing
Ana,
I was looking at Empresa.com before reaching out - your Meta Pixel is installed but not firing conversion events, which means you're building audiences in Meta without knowing which visits actually convert.
Worth a 15-minute call to show you what that's costing?
Carlos
Message 2 - Follow-up 1 (3-4 days after Message 1)
Purpose: Add value. Give something without asking for anything.
Format:
[First name],
[Reference to message 1 - don't apologize for following up]
[One piece of relevant information or insight - specific to their situation]
[Soft ask or no ask at all]
[Name]
Rules:
- Under 60 words
- Give value before asking again
- Can include a link to a relevant article or example (not your portfolio)
Message 3 - Follow-up 2 (5-7 days after Message 2)
Purpose: Create mild urgency or context. Last real attempt.
Format:
[First name],
[One line: what you've noticed since your last message - could be something about their industry, a competitor, or a seasonal angle]
[The ask again, slightly reframed]
[Name]
Rules:
- Under 50 words
- Do not apologize for following up
- Do not say "I know you're busy"
Message 4 - Break-up (7-10 days after Message 3)
Purpose: Close the loop, leave the door open. Often gets the highest reply rate.
Format:
[First name],
Closing this thread since I haven't heard back. No hard feelings - timing is everything.
[One line: what you'd have helped with]
If things change, you know where to find me.
[Name]
Rules:
- Under 40 words after the opening line
- No negativity, no guilt
- The "what you'd have helped with" line should be specific - it is your last pitch
References
references/email-frameworks.md- PAS, before/after, one-liner formatsreferences/personalization-signals.md- what to look for per company type
Output principles
- One specific observation per message - never generic
- Under 80 words for Message 1, shorter for each subsequent
- One ask per message - never a question AND a link
- No attachments in the sequence (unless asked)
- Personalization must be verifiable - only include things actually found in research
- Sequence must feel human, not automated - if it sounds like a mail merge, rewrite it
Gives 0 of the 12 instructions most sales crm skills give in ~1.4k tokens
Counted across 361 of the 361 authors here whose files we hold, read 2026-08-06
- Read product marketing context before writing if it existsin 22 of 361, across 14 files
- keep the ask low-frictionin 16 of 361, across 7 files
- Call RUBE_SEARCH_TOOLS firstin 15 of 361, across 5 files
- personalize every outbound messagein 13 of 361, across 4 files
- confirm connection status is activein 13 of 361, across 4 files
- Keep forwardable blurbs under 100 wordsin 13 of 361, across 4 files
- State if personalization context is missingin 13 of 361, across 4 files
- Cut any sentence that does not drive a replyin 13 of 361, across 4 files
- Use proof instead of adjectivesin 12 of 361, across 3 files
- Use a single, low-friction call to actionin 12 of 361, across 4 files
- Calibrate tone to the specific audiencein 12 of 361, across 3 files
- Make each follow-up email add new valuein 12 of 361, across 6 files
Said here and by no other author read
- research the prospect company and contact
- pick one primary pain angle
- connect observations to business outcomes
- keep subsequent messages shorter than message one
- include one specific observation per message
- ask one question or provide one offer
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.