# 32 Claude Skills for Outreach, From Who to Reach to the First Meeting 32 Claude skills that follow one outreach run in order, where Claude researches the company and the person with a source for every fact and you write and send every message. - Page: https://lantern.meet-polar-bear.com/skills/outreach-pack - Download: https://lantern.meet-polar-bear.com/skills/outreach-pack.zip - Source: https://github.com/polar-bear-org/claude-skills/tree/main/packs/outreach-pack - Price: free ## Install in Claude Code ``` /plugin marketplace add polar-bear-org/claude-skills /plugin install outreach-pack@polar-bear-skills ``` ## The 32 skills you get The set follows one outreach run in eight stages: deciding who to reach, researching the company, researching the person, finding the reason to write, writing it yourself, checking and sending in small runs, going from reply to meeting, and reviewing what worked. ### 1 · Decide who to reach - **Ideal Client Profile** — use when: You say yes to anyone and cannot say who a first message should go to. Output: Situations and triggers, the job you are hired for, roles in the decision, red flags, a test on a past engagement. - **Buying Situations Map** — use when: You know who you are for, not what would make them need you this month. Output: Moments that start a need, seen or assumed, the outside sign for each, the ones you leave alone. - **Target Account List** — use when: Your list is forty logos and no reason to write to any of them. Output: Two-line brief, criteria checked yes, no or unknown with sources, the number you can work, drops with reasons. - **Warm Path Check** — use when: You are about to write cold to a company where a former colleague works. Output: Who you already know per account from your own data, closeness in your words, a verdict per account. ### 2 · Research the company - **Company Fact List** — use when: You have twenty minutes and refuse to write from a guess. Output: Up to twelve numbered facts with date and link, claims dropped for no source, unknowns. - **Source Check** — use when: A fact sounds right and you cannot say where it came from. Output: Publisher, original or copy, date and wording checked per source, facts to keep, fix or drop. - **Company Research Brief** — use when: You have the facts and want to walk in with questions, not a pitch. Output: Half a page citing only numbered facts, hypotheses as questions, why now or "no reason yet". ### 3 · Research the person - **Contact Shortlist** — use when: You know the company and not who there would care. Output: Roles that live with the problem, people publicly named in them, the one or two you choose, the warm path. - **Identity Check** — use when: The search returns three people with the same name. Output: Each source tested for same name and same organisation, namesakes set aside, confirmed or not. - **Person Research Brief** — use when: You want to write something specific and refuse to guess or pad. Output: Sourced, dated facts about the person's work, what you share, unknowns, open questions. - **Research Coverage Check** — use when: The brief is three lines long and you are tempted to fill it. Output: Research labelled enough, thin or nothing found, what a thin result lets you say, your call. ### 4 · Find the reason to write - **Trigger Event Scan** — use when: You want a real reason to write now, not "just reaching out". Output: Recent public events with date and link, too old or too private set aside, both readings, pitch-hook test. - **Buying Signals Watchlist** — use when: The account fits and nothing has happened there yet. Output: One or two public signs per account, the page to watch, how often, what you do when one appears. - **Reason to Write** — use when: Every opening you think of turns into "let me tell you what I do". Output: Reasons that serve the other person, the fact each rests on, pitch-in-disguise test, one left for you to choose. - **Useful Material Match** — use when: You want to share something useful and everything feels like a brochure. Output: Your own materials checked against this person's facts: none, one or two that fit, and what not to send. ### 5 · Write it yourself - **Writing Voice Guide** — use when: Everything a tool drafts for you sounds like everyone else's. Output: How you open, ask and sign off, words you use and never use, an AI-tell checklist, a before and after. - **Talking Points** — use when: You want help thinking about what to say, not a machine-written message. Output: Two or three sourced points for the one reason you chose, no greeting, no assembled message. - **Cold Email Outline** — use when: You must write to someone who does not know you and refuse to send spam. Output: Warm path and past thread check, a frame with your words in each slot, flags per line, one follow-up date. - **Connection Request Note** — use when: A blank request will be ignored and a false "great to connect" is worse. Output: The true context per person, one specific detail, why connecting helps them, who comes off the list. - **Cold Call Outline** — use when: You avoid the phone because you do not know what to say after hello. Output: Keywords for the opening, the reason, two questions, the ask, two common refusals, a short voicemail. ### 6 · Check and send in small runs - **Message Checklist** — use when: You wrote it yourself and cannot tell whether it reads as one more pitch. Output: Six checks in order, a pass or a flag per check with the line, no replacement wording. - **Grammar Check** — use when: You want it clean without it stopping sounding like you. Output: Grammar, spelling and punctuation only, every change shown, risky fixes asked as questions. - **Outreach Rules Check** — use when: You write across borders and do not know which rule applies. Output: Channel, country and address type, the questions official pages raise, unknowns, questions for an adviser. - **Run of Ten Plan** — use when: Prospecting always comes last, then forty rushed notes go out. Output: About ten people this week, the time block, one reason and one status each, the next run's start. - **Follow-Up Tracker** — use when: Good conversations die because nobody wrote back. Output: Replies you owe first, one follow-up per silent thread, where each thread closes, calls to book. ### 7 · From reply to meeting - **Reply Triage** — use when: A reply arrives and you cannot tell interest from politeness. Output: What the reply says and asks, facts it adds, points for your answer, "stop" recorded at once. - **Conversation Opener** — use when: They accepted the connection and the next message is where most people pitch. Output: One reason, points from the real exchange, an ask sized to how little you know each other, an easy no. - **First Call Prep** — use when: They said yes to a call and you fear turning it into a pitch. Output: One page: why they agreed, three open questions, what to listen for, what not to pitch, a short agenda. - **Call Recap** — use when: A week later neither of you remembers what was agreed. Output: What they said, what each side promised and by when, the next step, points for your recap note. ### 8 · Review what worked - **Outreach Log** — use when: You cannot say who you wrote to last month or what you promised. Output: One private row per person: date, channel, reason, warm path, reply, call, open promises, do not contact. - **Outreach Review** — use when: You have no way to tell what worked. Output: Your own counts by reason, channel and warm or cold, expected against actual, one thing to change. - **Next Run Test** — use when: You keep changing five things at once and learn nothing. Output: One change, what you expect, what you count, how many runs before you look, keep or drop. ## Workflows - **Decide who to reach** (You cannot say who a first message should go to.) Ideal Client Profile → Buying Situations Map → Target Account List → Warm Path Check - **Research one company** (One account to work, and you refuse to write from a guess.) Company Fact List → Source Check → Company Research Brief - **Research the person** (You know the company and not who there would care.) Contact Shortlist → Identity Check → Person Research Brief → Research Coverage Check - **Find the reason** (You want a real reason to write now, not a pitch in disguise.) Trigger Event Scan → Reason to Write → Useful Material Match → Talking Points - **Write, check and send** (A first email to someone who does not know you.) Talking Points → Cold Email Outline → Message Checklist → Grammar Check → Outreach Rules Check → Run of Ten Plan - **From reply to call** (Replies and follow-ups are due, and one of them is a yes.) Follow-Up Tracker → Reply Triage → First Call Prep → Call Recap → Outreach Log - **Review what worked** (A few runs are done and you want to change one thing.) Outreach Log → Outreach Review → Next Run Test → Run of Ten Plan ## The quality bar - A source for every fact: every fact about a company or a person carries a link and a date, or it is marked as your own first-hand history. A fact with no source is dropped. Unknowns are written as unknown, and a thin result is labelled thin and stays thin - Warm before cold: the warm path is checked before any cold message, and a run holds about ten people - You write it: Claude gives a reason, points, a frame and flags. It never drafts the message, builds a sequence, places a call or sends anything. Where a browser or mail connector is used, it reads and does nothing else, and it is never used on LinkedIn - Organisations are checked, people are not: nothing grades, ranks, scores or profiles a person. Fit is checked for an organisation against your own brief, with yes, no or unknown - No invented numbers: no invented numbers, reply rates or benchmarks. Reviews use your own counts - Rules as questions: points about consent, anti-spam and data protection are questions taken from official pages. Check with a qualified adviser - The red line: Claude researches and prepares; you choose the people, write every word and press send: nothing is scraped, automated or sent for you, and a fact about someone with no source is dropped, never softened ## Questions people ask ### What is in the Claude Guide for Outreach pack? 32 Claude skills that follow one outreach run in eight stages: decide who to reach, research the company, research the person, find the reason to write, write it yourself, check and send in small runs, reply to meeting, and review. Each skill is one artifact. ### How do I install these Claude skills? In Claude Code, two lines install all 32 at once. In Claude.ai or Claude Desktop, download the zip, open Settings, Skills, Add, and upload the skills you want. Either way it takes about five minutes. ### Do I need Claude Code to use them? No. The skills work in Claude.ai, Claude Desktop and Claude Code. In the app you upload them yourself; in Claude Code one command installs the whole pack. ### Do the skills work with ChatGPT or other assistants? They are built as Claude Skills, and installing them is a Claude feature. Each one is a plain Markdown file, so you can read it and reuse the method elsewhere, but it will not load and trigger by itself outside Claude. ### Will Claude write or send the messages for me? No. Claude researches and prepares; you choose the people, write every word and press send. Nothing is scraped, automated or sent for you, and a fact about someone with no source is dropped, never softened. ### Is the pack free? Yes. It is free to use inside your company and in client work, and not for resale. There is no signup: the download and the install command are on this page. ## About Written by Polar Bear (https://meet-polar-bear.com/). Published on Lantern (https://lantern.meet-polar-bear.com/), which turns your notes, documents and call transcripts into social post drafts you review and publish.