agentleFS
camrontaylorcamrontaylor/AI-OS-Template/.claude/skills/mkt-icp/SKILL.md

mkt-icp

Skill0 starsChanged 45 days agoMITView on GitHub

Build or refine an ideal customer profile so every skill speaks to the right person. Triggers on: "target audience", "buyer persona", "ideal customer", "who am I selling to", "ICP", "customer avatar", "who buys this", "audience research". Two modes: Interview (build from founder knowledge) or Research (extract from URLs, reviews, social signals). Produces brand_context/icp.md. Foundation skill — run before execution skills that read audience context. Does NOT trigger for positioning, voice extraction, or content writing.

The file

Show its rules and text

We split this skill into the rules it gives an agent

17 rules. About 0 are common, 12 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈12

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If it exists → Update mode.

    line 37
  2. Distinctive:

    If it doesn't exist → Mode selection.

    line 39
  3. Distinctive:

    If the user provides a URL or mentions a specific platform in their opening message, skip to Research mode.

    line 46
  4. Very specific:

    Read references/interview-questions.md for the full question bank.

    line 54
  5. Distinctive:

    Ask a maximum of 8 questions.

    line 56
  6. Very specific:

    Read references/research-methods.md for the full process.

    line 66
  7. Distinctive:

    Run Interview mode first to build a hypothesis, then Research mode to validate and enrich with real customer language.

    line 81
  8. Distinctive:

    After gathering inputs from either mode, synthesise into the ICP profile.

    line 87
  9. Very specific:

    Read references/icp-template.md for the exact output format.

    line 89
  10. Distinctive:

    After building the profile, present a validation check:

    line 105
  11. Distinctive:

    Write 2 sentences as if speaking directly to this customer about their primary pain — one that should resonate, one that shouldn't.

    line 107
  12. Distinctive:

    Ask: *"Would your ideal customer read that first sentence and think 'this person gets me'?"*

    line 109
  13. Very specific:

    Write brand_context/icp.md using the format in references/icp-template.md.

    line 121
  14. Distinctive:

    If the file already existed, show what changed and confirm before overwriting.

    line 123
  15. Distinctive:

    After saving, show the user actual excerpts — the primary pain statement and 3-4 key phrases from their language section.

    line 125
  16. Very specific:

    If the user flags an issue with the output, add a dated entry under ## Rules in this skill's SKILL.local.md (create it if missing) and log …

    line 137
  17. Distinctive:

    Can't find customer language online: Ask the founder to share actual emails, DMs, or support tickets from customers.

    line 144

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

ICP — Ideal Customer Profile

Generic marketing talks to everyone and resonates with no one. A sharp ICP doesn't just describe demographics — it captures how the customer thinks, what words they use, what they've already tried, and why they're still stuck.

Outcome

brand_context/icp.md — a customer profile detailed enough that any skill can write copy, choose angles, or build content that feels like it was written by someone who knows the reader personally.

Context Needs

FileLoad levelHow it shapes this skill
brand_context/positioning.mdSummaryThe chosen angle tells us which audience segment to focus on
context/learnings.md## mkt-icp sectionApply previous insights about what resonates

Load if they exist. Proceed without them if not.


Before You Start

Check if brand_context/icp.md exists.

If it exists → Update mode. Read the file, show a one-paragraph summary, and ask what they'd like to change. Offer targeted updates (add a segment, refine language, update objections) rather than rebuilding from scratch.

If it doesn't exist → Mode selection. Ask:

"Do you already know who your ideal customer is, or should I research them?

  1. I know them — ask me questions and I'll describe them
  2. Research them — I'll pull signals from reviews, forums, and social media
  3. Both — interview me, then validate with research"

If the user provides a URL or mentions a specific platform in their opening message, skip to Research mode.


Mode 1: Interview

Best for: founders who know their customer but haven't documented the profile systematically.

Read references/interview-questions.md for the full question bank.

Ask a maximum of 8 questions. Prioritise based on what you already know from context. If positioning.md is loaded, skip questions it already answers.

Focus on surfacing their customer's own language — not marketing language. The words a customer uses to describe their problem are more valuable than how the founder frames it.


Mode 2: Research

Best for: validating assumptions, finding customer language, or starting without founder knowledge.

Read references/research-methods.md for the full process.

Sources to mine (in priority order):

  1. Customer reviews of the product or competitors (Amazon, G2, Capterra, app stores)
  2. Reddit/forum threads where the target audience discusses their problem
  3. Social media comments on competitor content
  4. YouTube comments on related videos
  5. Survey data or testimonials (if the user provides them)

The goal is to extract patterns in: how they describe the problem, what they've tried, what frustrated them, and what outcome they're after — in their exact words.


Mode 3: Interview + Research

Run Interview mode first to build a hypothesis, then Research mode to validate and enrich with real customer language. Flag any gaps between founder assumptions and actual customer signals.


Step 4: Build the Profile

After gathering inputs from either mode, synthesise into the ICP profile.

Read references/icp-template.md for the exact output format.

The profile must include all of these (each explained in the template):

  1. Who they are — not just demographics, but situation and identity
  2. Primary pain — the problem that keeps them up at night, in their words
  3. What they've tried — alternatives and why those didn't work
  4. Desired outcome — what "after" looks like for them
  5. Objections — why they hesitate, what makes them skeptical
  6. Their language — exact phrases, words, and framings they use
  7. Where they hang out — platforms, communities, publications they trust
  8. Buying triggers — what pushes them from "considering" to "buying"

Step 5: Validate

After building the profile, present a validation check:

Write 2 sentences as if speaking directly to this customer about their primary pain — one that should resonate, one that shouldn't.

Ask: "Would your ideal customer read that first sentence and think 'this person gets me'?"

  • Yes → save
  • Close but off → ask what feels wrong, adjust, retest
  • Not right → dig deeper on language and pain

Cap at 3 rounds.


Step 6: Save Output

Write brand_context/icp.md using the format in references/icp-template.md.

If the file already existed, show what changed and confirm before overwriting.

After saving, show the user actual excerpts — the primary pain statement and 3-4 key phrases from their language section.


Rules

Updated automatically when the user flags issues. Read before every run.


Self-Update

If the user flags an issue with the output, add a dated entry under ## Rules in this skill's SKILL.local.md (create it if missing) and log it to context/learnings.md under ## mkt-icp. Write the general rule the correction implies, not the one-off fix.


Troubleshooting

Founder describes the customer too broadly: Push for specifics. "Entrepreneurs" is too broad. "Solo founders making $5-15k/month who do their own marketing" is useful. Can't find customer language online: Ask the founder to share actual emails, DMs, or support tickets from customers. Multiple distinct segments emerge: Build the primary ICP for the most valuable segment. Note secondary segments at the bottom of the file for future reference. Positioning not loaded: Proceed, but note that the positioning angle would help prioritise which segment to focus on. ICP and positioning conflict: Flag it. The ICP might reveal the positioning angle needs adjustment — that's valuable signal.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.