cold-outreach-writer
Skill3 starsChanged 6 months agoMITView on GitHub
Write a personalised cold outreach email from a company research block. Produces one tight email under 180 words built on a specific observation about the prospect, plus three alternative subject lines, a LinkedIn connection note, and a 7-day follow-up. Drafts only, never sends anything. Use when the user says "write a cold email", "outreach email for this lead", "draft outreach for this company", or "follow up email for this prospect".
1 other project uses a version of this.
Adopted by
Most recently taken up by
- landedjobs/ai-job-hunt-ostheir own version, first seen Jul 24, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- landedjobs/ai-job-hunt-os★ 3same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository47 other files it gives its agents.
- agents-md-writerSkill
- ai-image-run-pacingSkill
- automation-state-reconcilerSkill
- backup-readiness-auditorSkill
- brand-voice-linterSkill
- case-study-writerSkill
- client-delivery-auditorSkill
- client-intake-builderSkill
- client-workspace-scaffolderSkill
- code-deduplicatorSkill
- demo-animationSkill
- demo-animation-v2Skill
- discovery-briefSkill
- em-dash-removerSkill
- etsy-listing-optimizerSkill
- follow-up-radarSkill
- github-profile-proof-updaterSkill
- github-safe-publisherSkill
- idea-inboxSkill
- implement-client-feedbackSkill
- lead-scorerSkill
- llms-txt-generatorSkill
- local-app-public-demoSkill
- media-spec-checkerSkill
- obsidian-daily-noteSkill
- ops-status-reportSkill
- outreach-pipeline-designerSkill
- php-form-mailerSkill
- proposal-writerSkill
- release-announcement-writerSkill
- remove-ai-image-metadataSkill
- research-brief-curatorSkill
- research-to-articleSkill
- screen-recorderSkill
- session-handoffSkill
- skill-builderSkill
- skill-reviewerSkill
- social-campaign-proof-trackerSkill
- skill-nameSkill
- thumbnail-legibility-checkSkill
- thumbnail-set-designerSkill
- vault-janitorSkill
- verify-public-claimsSkill
- web-scraperSkill
- website-change-verifierSkill
- windows-reinstall-plannerSkill
- write-evidence-backed-profileSkill
The file
We split this skill into the rules it gives an agent
9 rules. About 0 are common, 8 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Look for an outreach-profile.md file in the project root.
- Distinctive: line 25
Never invent proof points, client names, or numbers.
- Distinctive: line 29
Ask for (or accept as input) a research block about the prospect:
- Distinctive: line 48
Only use research the user provides or public information they point you to.
- Distinctive: line 52
Write ONE cold email with these rules:
- Distinctive: line 84
Never fabricate: Every claim must come from the profile or the research block.
- Distinctive: line 99
No em dashes in any output.
- Distinctive: line 100
Do not scrape private data or guess email addresses.
- Distinctive: line 101
If the user asks for volume ("write 50 of these"), push back once: suggest writing 3 strong ones and reviewing results first.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Cold Outreach Writer
You write cold outreach emails that get replies because they prove the sender actually looked at the prospect's business. One specific observation beats ten lines of flattery. This skill produces drafts for human review. It never sends anything.
Step 0: Load the sender profile (never guess these)
Look for an outreach-profile.md file in the project root. If it does not exist, ask the user for the values below, then offer to save them to outreach-profile.md for next time.
NAME Who the email is from
TITLE Positioning title, e.g. "Brand Identity Specialist", not a generic job label
WEBSITE Public site to link in the signature
PROOF_POINTS 2 or 3 real, verifiable numbers, e.g. "300+ projects, 15 countries, 10 years"
SERVICES What the sender actually sells, in their own words
PORTFOLIO Industry to example mapping, e.g. "Real estate: [project name and one-line result]"
TONE_NOTES Optional voice guidance, e.g. "peer-to-peer, selective, never desperate"
EXTRA_BANNED Optional additional words or phrases to never useNever invent proof points, client names, or numbers. If a field is missing, write the email without it rather than fabricating.
Step 1: Collect the company research block
Ask for (or accept as input) a research block about the prospect:
Company Name:
Industry:
Location (City, Country):
Website:
Decision Maker Name:
Decision Maker Title:
Website content summary (what their site says and how it looks):
Brand/business weaknesses observed:
Growth signals (new office, expansion, hiring, recent news):
Most relevant portfolio example for this industry (from the profile):Only use research the user provides or public information they point you to. If the research block is thin, say so and ask for the missing pieces. A cold email built on guesses reads like spam because it is.
Step 2: Write the email
Write ONE cold email with these rules:
- Subject line: specific and intriguing, never salesy. Reference the company name plus one real observation. Patterns that work:
- "[Company]'s brand: one thought"
- "A thought on [Company]'s visual direction"
- "[Company]: what I'd do differently"
- Opening (1 sentence): a SPECIFIC observation about their gap, not flattery. Base it on the research block.
- BAD: "I appreciate your impressive work in the market."
- GOOD: "Your properties speak premium, your current brand identity doesn't quite match that ambition yet."
- Credibility (1 sentence): one relevant portfolio example or proof point from the profile, same or adjacent industry. Skip it if the profile has nothing relevant; an unrelated example is worse than none.
- Offer (1 sentence): what the sender brings, framed as ideas and value, not "services available".
- CTA (1 sentence): low friction, no commitment language. Example: "Would you be open to me sharing a quick visual breakdown?"
- Signature: NAME, TITLE, WEBSITE from the profile.
- Length: MAXIMUM 180 words. Every word must earn its place.
- Tone: confident, direct, peer-to-peer. The sender is busy and selective, and the email reflects that. Apply TONE_NOTES if provided.
- Banned phrases (plus anything in EXTRA_BANNED):
- "I hope this email finds you well"
- "just checking in"
- "excited about the opportunity"
- "I am available" / "free to hop on a call"
- "overflow work" / "assist your team"
- "full-service" / "one-stop shop"
- "look no further"
- Never fabricate. Every claim must come from the profile or the research block. No invented mutual connections, no fake urgency, no made-up compliments.
Step 3: Deliver the full package
Return structured text with clear section headers:
- The email (subject plus body)
- 3 alternative subject lines
- LinkedIn connection note for the same prospect (300 characters max)
- Follow-up email for 7 days later if no reply (80 words max, references the first email in one clause, adds one new angle, does not guilt-trip)
Rules
- Drafts only. Never send email, connect to an email service, or automate delivery. A human reads and sends every message.
- One prospect per run. Personalisation does not batch.
- No em dashes in any output. Use commas, colons, full stops, or brackets.
- Do not scrape private data or guess email addresses. Research comes from the user.
- If the user asks for volume ("write 50 of these"), push back once: suggest writing 3 strong ones and reviewing results first.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

