Claude marketing skills: catalog, use, and build

Claude marketing skills are reusable job packages - a SKILL.md plus optional references - not a chat prompt dump and not “install everything marketing.” Match skill types to weekly jobs (positioning, voice, briefs, SEO pages, email, social, QA), load brand inputs in a Project first, then install or author a skill when the loop is stable. For “what to install first,” use What Claude marketing skills to install first. Cluster map: Claude for marketing. Prompt pack before skills: Claude prompts for marketing.

When to use Claude marketing skills / when not

Use skills when you repeat a marketing job weekly with stable inputs: ICP notes, voice examples, offer facts, proof you can stand behind, and a clear QA bar. Same brief shape every run - skill territory.

Don’t use skills yet on day zero with empty brand docs, one-off spikes, strategy/taste calls, or a “40-skill team pack.” Skills don’t invent positioning you refuse to write. Overlapping descriptions compete for trigger space - more skills can make the right one less likely to fire.

If you only need install order and anti-bloat criteria, stop here and open What Claude marketing skills to install first. Day-to-day Projects loop: How to use Claude for marketing.

Catalog map - skill types by marketing job

Catalog by job, never by unverified marketplace SKU names. Names on plugin browsers change; verify same-day if you install a third-party skill. The rows below are types, not product titles.

Do not invent live marketplace titles here. Parallel SEO path: What Claude SEO skills to install first. Model/tool choice for marketing jobs: Claude vs ChatGPT for marketing.

Workflow - use or build in 8 steps

1. Inventory weekly jobs - List what you actually run (≥1×/week). Start from the marketing map and the Projects how-to if the list is fuzzy.
2. Load a Project brand pack - Voice rules, 2-3 examples, offer + ICP, optional positioning one-pager. Skills without context still produce generic output.
3. Run the job as a prompt until the loop is boring - Spec role · inputs · steps · rules · output. Full cards: Claude prompts for marketing.
4. Choose the surface - Chat for one-offs; Projects for brand context + repeat briefs; Claude Code for repo workflows and commands; marketplace/plugins only after same-day name/description check.
5. Install or author - Matching plugin/marketplace skill or a thin custom SKILL.md when the prompt barely changes between runs.
6. Test the trigger description - Ask a teammate to phrase the job three ways; the description must match what people actually say.
7. QA the output - No invented metrics, testimonials, or competitor facts; one job per run; human gate for customer-facing copy.
8. Prune - Remove unused skills. Keep the install-first spirit: few skills that earn their keep, not a directory dump.

Prompts (≥3 copy-paste)

Three specs for skill decisions - not the five-card marketing pack. Mirror: prompts.md. Depth on positioning, brief, voice, section draft: Claude prompts for marketing.

1. Skill-need diagnostic

Role: Marketing ops lead. Inputs: job list · frequency · brand inputs available · current prompt stability. Output: skill / prompt / defer table + rationale. When not: day-zero with zero brand docs.

You are a marketing ops lead. Decide which jobs deserve a Claude skill vs stay as prompts.

REQUIRED INPUTS: Weekly job list (name + frequency); Brand inputs available (voice/ICP/proof or "none"); For each job: prompt stable? (yes/no/unknown); Surface preference (chat|Projects|Claude Code|marketplace|unknown).

STEPS: (1) Score each job: repeat weekly? has inputs? stable loop? clear QA bar? (2) Table: Job | Verdict (skill|prompt|defer) | Why | Next action. (3) Cap "skill" recommendations at 3 unless every row clearly passes. (4) Never invent marketplace skill product names  -  use job-type labels only. (5) Route "what to install first" to the install-order companion, not a ranked SKU list.

RULES: Fail any score row  -  prompt or defer. No "install everything marketing" advice. Mark [GAP] where inputs are missing.

OUTPUT: decision table + 3 next actions + QA self-check (no invented skill names, cap respected).

2. SKILL.md draft from a stable prompt

Role: Skill author. Inputs: winning prompt · job name · trigger phrases · optional refs. Output: frontmatter + instructions body. When not: prompt still changing weekly.

You are a Claude Agent Skills author. Convert one stable marketing prompt into a thin SKILL.md.

REQUIRED INPUTS: Job name (kebab-case candidate); Winning prompt (full text); When to use (3-5 user phrasings); When not; Required inputs list; Optional reference filenames (or "none"); Surface (claude.ai|Claude Code|both).

STEPS: (1) Draft YAML frontmatter with name (lowercase kebab-case) and description that states WHAT + WHEN. (2) Body: Instructions (steps), Required inputs, Output format, QA checklist, When not. (3) Keep body lean; put long examples in optional refs. (4) Flag any hard-coded product facts that will expire. (5) Do not add marketplace packaging claims.

RULES: Description ≤1024 chars; name ≤64 chars, a-z/0-9/hyphens only. No invented metrics in examples. Label stub vs production if packaging is incomplete.

OUTPUT: complete SKILL.md in a fenced block + trigger-test checklist (3 alternate phrasings).

3. Trigger / description QA

Role: Skills QA. Inputs: draft description · sibling skill descriptions · sample requests. Output: pass/fail + rewrite. When not: no sibling list and no real teammate phrasing.

You are QA for Claude marketing skills. Stress-test a skill description for false triggers and misses.

REQUIRED INPUTS: Candidate skill name; Draft description; Sibling skill descriptions (or "none"); 5 real request phrasings from the team; Jobs this skill must NOT own.

STEPS: (1) For each phrasing: trigger? (yes/no/borderline) + why. (2) Flag overlap with siblings. (3) Rewrite description if misses or overlaps. (4) List phrases that should stay on prompts, not this skill. (5) Never invent competing marketplace skill titles.

RULES: Prefer precise WHEN language over buzzwords. One job per skill. Cap rewrite at 1024 characters.

OUTPUT: trigger table | overlap notes | revised description | ship/hold verdict.

Agent block - Marketing skill librarian

Paste into Project instructions or a Claude Code custom agent. Enforces job match, no invented marketplace names, and routes install-order questions to the companion guide.

You are the Marketing skill librarian for this brand.

RULES:
1. Map every request  -  marketing job type (positioning, voice, campaign brief, SEO brief, email, social/repurpose, claims/QA, other).
2. Recommend skill OR prompt OR defer  -  never invent marketplace/plugin skill product names. Use job-type labels only.
3. If brand inputs (voice, ICP, proof) are missing, mark [GAP] and ask  -  do not invent fills.
4. Install-order / "what should I install first"  -  point to https://claude-for.com/guides/what-claude-marketing-skills-to-install-first ; do not invent a day-one top-3 SKU list.
5. Prefer Projects + prompts until the loop is stable; graduate to SKILL.md only when the prompt barely changes.
6. Customer-facing output is always draft until a human QA gate.

OUTPUT FORMAT each turn:
- Job type
- Recommendation (skill type | prompt | defer)
- Next action (1 sentence)
- QA checklist (inputs present? one job? no invented metrics? human gate?)

Skills path / stub

Honest stub in this pack: skill-stub/SKILL.md (marketing-skill-catalog navigator). Structure only - not a downloadable zip and not a marketplace listing.

Surfaces (verify same-day against official docs):

General skills primer (any domain): Claude skills. Anatomy (required): YAML name + description (what and when), then Markdown instructions; optional refs/scripts load on demand. Official overview: Agent Skills · Claude Code: Extend Claude with skills · (Checked 2026-09-14).

SEO track parallel: What Claude SEO skills to install first.

Stack glance

Projects beat skills when context is the bottleneck. Skills beat re-pasting when the procedure is stable. Claude Code wins when the marketing job lives next to a repo (site content, skill eval, command wrappers). Comparison spoke if the question is Claude vs ChatGPT for the job: Claude vs ChatGPT for marketing.

Pitfalls + quality bar

• Skill bloat - overlapping descriptions dilute triggers; prune aggressively.
• Rules vs examples contradiction - voice skill that bans slang while examples use slang.
• Hard-coded expired product facts - put volatile claims in Project files, not the skill body.
• Listicles as product truth - third-party “best marketing skills” posts are not Anthropic docs.
• Skipping install-order discipline - this page is catalog/use/build; day-one caps live on install-first.

Quality bar: one job · sourced proof · on-brand vs examples · no invented metrics · human QA for customer-facing.

FAQ

What are Claude marketing skills?
Reusable packages (SKILL.md + optional refs) that teach Claude a marketing job and when to load it. Not the same as a one-shot chat prompt, a full agent, or an MCP server. Official definition: modular capabilities with progressive disclosure - metadata always loaded, body when triggered (Agent Skills overview, Checked 2026-09-14).

How are they different from prompts?
A prompt is conversation-level for one run. A skill is filesystem-based: Claude matches the description, then loads instructions automatically. Stabilize prompts first - pack: Claude prompts for marketing.

Should I install skills before writing prompts?
Usually no. Projects + prompts until the loop stops changing, then graduate. Install order and anti-bloat: What Claude marketing skills to install first.

How do I build a marketing SKILL.md?
Directory with SKILL.md: frontmatter name + description (what and when), Markdown instructions, optional refs. Test on a live brief with teammate phrasing. Stub outline: skill-stub/SKILL.md.

Claude Code vs Projects for marketing skills?
Projects for brand context and repeat briefs in claude.ai. Claude Code when skills live in a repo, need / commands, plugins, or agent/eval loops. Custom skills do not sync across surfaces automatically - upload or place them per surface (Checked 2026-09-14).

Key takeaways

• Catalog Claude marketing skills by job type, not marketplace fame.
• Load brand inputs in a Project; graduate stable prompts into thin SKILL.md files.
• Use chat · Projects · Claude Code · marketplace deliberately; verify plugin names same-day.
• Cap early installs; prune unused skills; never invent skill product titles in docs.
• Install order - install-first guide; loop - how-to; map - Claude for marketing; prompts - prompts pack.

Sources (checked 2026-09-14)

• Agent Skills overview - anatomy, progressive disclosure, surfaces
• Extend Claude with skills (Claude Code) - filesystem paths, /skill-doctor, plugins
• Equipping agents for the real world with Agent Skills - engineering background
• Live cluster: Claude for marketing · How to use Claude for marketing · Claude prompts for marketing · What Claude marketing skills to install first · What Claude SEO skills to install first · Claude vs ChatGPT for marketing
• Hub claude-for-growth - planned; not linked (404 as of 2026-09-14)