What these skills are
The junior marketing job, as installable Claude skills.
This is Claude for the work a marketing graduate is actually handed: the brand's voice written down, the audience read from what customers say, the campaign planned and launched, the content made across channels, the numbers reported in plain words — and then the proof project, CV and interview answers that turn all of it into something an employer can read.
The skills follow the job in the order the work arrives, in seven stages, and each stage hands a named artifact to the next. Setting up AI-fluent work hands over the brand voice guide, the prompt briefs, the draft review and the claim check. Research hands over the personas, the jobs map, the competitor map and the SWOT. Planning hands over the creative brief, the messaging house, the campaign plan, the See Think Do Care plan and the launch checklist. Creating hands over the calendar, the posts, the video script, the newsletter, the landing page and the SEO brief. Reporting hands over the metrics readout, the monthly report and the test plan. Then the proof project hands over a case study, an honest AI note and a portfolio post, and the last stage hands over the decoded ad, the CV, the letter, the task prep and the interview practice.
The red line: Claude helps you prepare, practise, research, organise and critique your own work. It never sits a test, interview or assessment for you, and it never invents experience.
Each skill is one named method or one artifact, with the inputs it needs, where it breaks, a fixed output template, a finish line and the next skill to run. In marketing work nothing is invented: no made-up statistic, result, review, testimonial or customer quote — every claim in your copy traces to evidence someone holds, and anything without a source is marked as an assumption or left as a bracketed placeholder.
Where a skill fits a Claude surface — Projects to keep the voice guide and briefs in one place, Claude in Chrome to read public pages for research, Claude Docs for plans and CVs, Claude Slides for reports and portfolio decks — it says so, and beta features are marked beta. Every skill also works in a normal chat with what you paste in.
Mechanically, each skill is one folder with a SKILL.md file. The 31 are grouped into the seven stages, and each one gives full value on its own: run one, run a stage, or work the job through in the order it arrives.
Download all 31 skills. One zip: ready-to-install skill zips, readable SKILL.md files, and the sources behind every method. Free, no signup. The same 31 skills are open on GitHub: polar-bear-org/claude-skills · Download the zip
The 31 skills you get
The set follows the junior marketing job, in seven stages: setting up AI-fluent work, researching the audience and market, planning the campaign, creating content in the brand voice, reporting and improving — then building a proof project and applying for the job.
1 · Set up AI-fluent marketing work
1. Brand Voice Guide
Use when: You write for a brand that is not yours and every draft sounds like a different person
Output: “We are / we are not” attributes, tone by situation, word list, three before-and-after rewrites, Project instructions block
2. Marketing Prompt Brief
Use when: Your AI drafts come back generic because the prompt was one line
Output: Reusable brief per recurring task, three starters, what never goes into a prompt
3. AI Draft Review
Use when: The draft reads fine and nobody has actually checked it
Output: Line-by-line check for facts, voice, generic tells and unsourced numbers, kept / changed / cut log, sign-off line
4. Claim Substantiation Check
Use when: The copy says “best” or a percentage and nobody knows where it came from
Output: Every claim listed, evidence held or missing, reviews and ad labels checked, rewrite or cut, questions for an adviser
2 · Research the audience and market
5. Audience Persona
Use when: The brief says “millennials” and nobody has read what customers write
Output: One or two personas from public evidence, each line cited, assumptions marked, three things to test
6. Jobs to Be Done Map
Use when: You know who buys but not why they switched or hesitate
Output: Job statements, push, pull, anxiety and habit from real reviews, the message each force suggests
7. Competitor Content Audit
Use when: Your manager asks what competitors are doing on social by Friday
Output: Channels, themes, formats and rhythm, claims sorted into table stakes, rare and specific, gaps; a map, never a ranking
8. SWOT Analysis
Use when: Your SWOT is four lists with no “so what”
Output: SWOT with a source per line, PESTLE prompts, paired factors turned into actions
3 · Plan the campaign
9. Creative Brief
Use when: The designer or agency needs a brief and you have a meeting note
Output: Objective, audience, single-minded proposition, reason to believe, mandatories, deliverables, measures, deadline
10. Messaging House
Use when: Every channel says something slightly different about the same launch
Output: Umbrella message, three pillars, sourced proof points, words to avoid, one line per channel
11. Campaign Plan
Use when: You are asked to “pull together a plan” for a launch or event
Output: Situation, SMART objectives with a baseline, strategy, tactics by channel, actions with owners, control
12. See Think Do Care Plan
Use when: All the content talks to people ready to buy and nobody else
Output: Audience at each stage, content and channel per stage, one metric per stage, gaps in the mix
13. Campaign Launch Checklist
Use when: Launch is tomorrow and the links, tags and approvals are unchecked
Output: UTM naming sheet, link and tracking tests, named approvals, ad labels, alt text, go or no-go owner
4 · Create content in the brand voice
14. Content Calendar
Use when: You post when there is time and the month has no shape
Output: Month grid by channel, pillar, format, owner, status and approval date, evidence behind each post
15. Social Media Posts
Use when: You need this week's posts in the brand voice and they all sound like AI
Output: One idea as LinkedIn, Instagram and TikTok variants, three hooks, alt text, “Ad” label where needed
16. Short Video Script
Use when: The team wants a reel and you have a phone and an afternoon
Output: Opening hook, beat-by-beat script, shot list, on-screen text and captions, edit notes
17. Email Newsletter
Use when: The newsletter is due and last month's got few opens
Output: Three subject lines with preview text, body, one call to action, plain-text version, a subject line test
18. Landing Page Copy
Use when: A campaign needs a page and you have a product sheet
Output: Hero line, problem, benefits, sourced proof, objections and FAQ, call to action, claim list
19. SEO Content Brief
Use when: You are asked for “a blog post for SEO” and have a topic, not a query
Output: Target query, search intent, outline from real questions, title and meta description, internal links, draft
5 · Report and improve
20. Marketing Metrics Readout
Use when: You have a dashboard full of numbers and do not know which matter
Output: Each metric defined in plain words, change against last period, what it cannot tell you, questions before reporting
21. Monthly Marketing Report
Use when: The monthly report is due Monday and last month's was a screenshot dump
Output: Bottom line first, results against objectives, what worked with evidence, next month's actions, the decision
22. A/B Test Plan
Use when: Someone asks what to test next and the answer is a hunch
Output: Test ideas ordered by ICE, one hypothesis, one variable, metric, duration, decision rule written first, result log
6 · Build a proof project
23. Proof Project Plan
Use when: Every ad wants marketing experience and you have none to show
Output: A real brand from public information, a two-week scope, which skills to run, what you will not claim, sources
24. Marketing Portfolio Case Study
Use when: You did the project and need it readable in two minutes
Output: Brief, insight, idea, what you made, what you would measure, what you learned; one page and a slide version
25. How I Used AI Note
Use when: You used AI on the project and do not know how to say it
Output: Tasks you gave Claude, briefs used, what you checked and changed, what you did yourself, a three-line version
26. Portfolio LinkedIn Post
Use when: The project is done and you want people hiring in marketing to see it
Output: One moment from the project in your words, what you made, the AI note in one line, no brand tagged as a client
7 · Apply and interview
27. Job Ad Decoder
Use when: Twenty marketing ads look the same and you cannot tell which you can evidence
Output: Tasks and tools in the ad, must-have against nice-to-have, your real evidence per line, gaps and how to close them
28. Marketing CV
Use when: Your CV says “social media” and “Canva” and nothing an ad asks for
Output: Bullets as task, tool and outcome from real experience, a projects section, a true tools line, two-page layout
29. Marketing Cover Letter
Use when: Every letter you write could go to any brand
Output: Why this brand with two specifics, one proof story, an honest line on AI, in your own voice
30. Marketing Task Prep
Use when: You have a take-home task or case and freeze at the blank page
Output: Task brief decoded, plan and timing, practice on a sample brief, critique of your own draft, rehearsal questions
31. Marketing Interview Questions
Use when: The interview is Thursday and you have never critiqued a brand out loud
Output: Likely questions from the ad, STAR stories from real experience, a mock interview, feedback on your answers
Where to start, and how the skills chain
There is no compulsory sequence, but there is one good first move: if you write for a brand, run the Brand Voice Guide, because every content skill after it checks drafts against what it produces. Otherwise bring whatever is due — the newsletter, the monthly report, a competitor question by Friday, a take-home task — and run the one skill named for it. Every skill lists the inputs it needs, so it tells you when the stage before it is the one that is actually missing.
They chain. gmkt-brand-voice-guide and gmkt-prompt-brief set up the work that gmkt-ai-draft-review and gmkt-claim-check check before anything is sent. gmkt-audience-persona and gmkt-jobs-to-be-done read the audience out of public evidence, with gmkt-competitor-content-audit and gmkt-swot-analysis mapping the space around it. gmkt-creative-brief and gmkt-messaging-house settle what the work says, gmkt-campaign-plan and gmkt-see-think-do-care sequence it, and gmkt-launch-checklist clears it for launch. gmkt-content-calendar gives the month a shape that gmkt-social-media-posts, gmkt-short-video-script, gmkt-email-newsletter, gmkt-landing-page-copy and gmkt-seo-content-brief fill. gmkt-metrics-readout explains the numbers, gmkt-monthly-report reports them and gmkt-ab-test-plan turns what did not work into the next test. Then gmkt-proof-project-plan builds experience you can show, gmkt-portfolio-case-study, gmkt-how-i-used-ai and gmkt-portfolio-post write it up, and gmkt-job-ad-decoder, gmkt-marketing-cv, gmkt-marketing-cover-letter, gmkt-marketing-task-prep and gmkt-interview-questions take it to the application.
The sources behind each method are in resources/evidence-and-sources.md, inside the pack.
Setup guide
- Download the pack. One zip: an install folder with 31 ready-to-upload skill zips, a skills folder with the same 31 skills as readable SKILL.md files, and resources/evidence-and-sources.md, which names the source behind every method and where it stops.
- Install your skills. In Claude Code, add the marketplace and install grad-marketing-pack, and all 31 load at once. In Claude, turn on code execution in Settings, Capabilities, then go to Customize, Skills and upload one zip per skill from the install folder. Prefer working from files? Add the SKILL.md files to your Project knowledge instead; it works, just less cleanly.
- Start with the brand voice, or with whatever is due. If you write for a brand, run the Brand Voice Guide first: every other content skill checks drafts against it. If something is due this week — the newsletter, the monthly report, a take-home task — run the one skill named for it. Each skill lists the inputs it needs and names the next skill to run, so one run turns into the chain on its own.
Where to start
| Your situation | Skill to run |
|---|---|
| Every draft sounds like a different person? | Brand Voice Guide |
| The AI draft reads fine and nobody has checked it? | AI Draft Review |
| The copy says “best” and nobody knows where it came from? | Claim Substantiation Check |
| The brief says “millennials” and that is all? | Audience Persona |
| Your manager wants to know what competitors post, by Friday? | Competitor Content Audit |
| You are asked to pull together a plan for a launch? | Campaign Plan |
| This week's posts are due and they all sound like AI? | Social Media Posts |
| A dashboard full of numbers and no idea which matter? | Marketing Metrics Readout |
| Every ad wants experience you do not have yet? | Proof Project Plan |
| The interview is Thursday? | Marketing Interview Questions |
The quality bar
Every skill in the pack holds the same standard, the one we hold when we market our own work:
- One method, applied properly: each skill uses the real mechanics of its source — the tone dimensions, the SOSTAC stages, the See Think Do Care stages, the UTM fields, the GA4 definitions — not a generic gather, analyse, recommend
- Evidence before claims: every number, proof point, persona line and competitor claim comes with where it came from, and anything without a source is marked as an assumption or left as a bracketed placeholder
- A person checks every draft: AI drafts go through a review that records what a person checked, kept, changed and cut before anything is sent or published
- The brand's voice, your own voice: brand copy follows the brand's voice guide, while applications, posts and interview answers stay in your words, from your own experience
- Work is scored, people are not: skills score evidence, options, test ideas and fit to a brief, never a person, a competitor's staff or a candidate — and advertising, consumer law and data protection points end with “check with a qualified adviser”
- The red line: Claude helps you prepare, practise, research, organise and critique your own work. It never sits a test, interview or assessment for you, never invents experience, and never writes assessed work unless the rules allow it and you say so
- Nothing invented in marketing copy: no made-up statistic, result, review, testimonial or customer quote — every claim traces to evidence someone holds
Who made this
Polar Bear is a people ops consultancy for human-size teams (20 to 200 people). Built by ex-McKinsey founders with a dream to make AI work for People, not instead of them. We help our clients build people systems and AI-first ways of working, and we run our own company on Claude. This pack is the free, self-serve version of how we work.
The pack carries one junior marketer's whole job. When you want your whole team working this way, AI carrying the overhead so people do the part only people can do, across marketing, hiring and everyday operations, that's what we build with clients.
Meet Pauline. Starting your marketing career out of one-off chats? Bring the stage you are stuck on and we will pick the skills for it. Book a 30-minute call · Pauline on LinkedIn
Install in Claude Code
Two lines, and every skill in the pack loads at once.
/plugin marketplace add polar-bear-org/claude-skills /plugin install grad-marketing-pack@polar-bear-skills
Using Claude on the web instead? Download the zip and upload each skill from its install folder under Customize → Skills.