agentleFS
adidactaadidacta/pmf-detective/skills/plan-pmf-mode/SKILL.md

plan-pmf-mode

Skill18 starsChanged 9 months agoMITView on GitHub

Guide users through building their complete PMF context layer. Triggered by /plan-pmf command or when user wants to build full context.

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. About 0 are common, 10 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈10

    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:

    Follow the full icp-builder flow: Do NOT ask inline ICP questions here — the icp-builder defines all 5 phases:

    line 54
  2. Very specific:

    When the icp-builder flow completes and saves pmf/icp.md, show the transition:

    line 62
  3. Distinctive:

    Follow the full value-prop-builder flow: Do NOT ask inline value prop questions here — the value-prop-builder defines all 5 phases:

    line 93
  4. Very specific:

    When the value-prop-builder flow completes and saves pmf/value-prop.md, show the transition:

    line 101
  5. Distinctive:

    Follow the full mvp-builder flow: Do NOT ask inline questions here — the mvp-builder defines all 7 phases:

    line 133
  6. Very specific:

    When the mvp-builder flow completes and saves pmf/mvp.md, show the transition:

    line 143
  7. Very specific:

    When the validation-plan-builder flow completes and saves pmf/validation-plan.md, show the transition:

    line 179
  8. Distinctive:

    If Landing page: Follow the landing-generator flow.

    line 202
  9. Distinctive:

    If Conversations: Follow the outreach-builder flow.

    line 203
  10. Distinctive:

    If Build & test: Follow the build-test-guide flow.

    line 204
  11. Distinctive:

    Ask ONE question at a time

    line 243
  12. Distinctive:

    Never skip sections without user consent

    line 245
  13. Distinctive:

    Keep headers to 1 word max or omit entirely

    line 256
  14. Distinctive:

    Don't rely on text above the question to provide context

    line 257
  15. Very specific:

    All outputs go to the pmf/ folder:

    line 261

≈ “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

PMF Plan Mode - Full Context Builder

You guide product builders through creating their complete PMF context layer in sequence: ICP → Value Proposition → MVP PRD → Validation Plan → Execute.

Important: The first 3 sections are assumptions. The validation plan decides how to test them, then routes to the right execution skill.

How This Skill Works

This skill orchestrates the full flow by following each builder's process in sequence. Each builder skill (icp-builder, value-prop-builder, etc.) defines its own phases and questions — follow their SKILL.md instructions directly. This skill provides the sequencing, transitions, and progress tracking between sections.

Your Role

  • Walk users through all sections in order
  • For each section, follow the corresponding builder skill's full flow
  • One question at a time
  • Save each section when complete
  • Move to next section automatically

The Flow

┌───────────────────────────────────────────────────────────────┐
│  PMF CONTEXT BUILDER                                          │
├───────────────────────────────────────────────────────────────┤
│                                                               │
│  Let's build your PMF context layer.                          │
│                                                               │
│  We'll work through 4 steps + execution:                      │
│  1. ICP — Who your customer is (assumption)                   │
│  2. Value Proposition — Why they should care (assumption)     │
│  3. MVP PRD — What to build & why (assumption)                │
│  4. Validation Plan — How you'll test these assumptions       │
│     │                                                         │
│     ├─→ Landing Page (if validating with signups)             │
│     ├─→ Outreach + Mom Test (if validating with talks)        │
│     └─→ Build with BMAD (if validating by building)           │
│                                                               │
│  Each section becomes a reference file Claude uses            │
│  when building anything for your product.                     │
│                                                               │
└───────────────────────────────────────────────────────────────┘

Section 1: ICP (icp-builder)

Follow the full icp-builder flow. Do NOT ask inline ICP questions here — the icp-builder defines all 5 phases:

  1. Context & Broad Target (2 questions)
  2. Pain Discovery with 5 Whys (3-6 questions)
  3. Hypothesis Formation (2-3 questions, produces 3 hypotheses)
  4. Parallel Research (automated — launches 3 research agents)
  5. Compare & Select (user picks the strongest hypothesis)

When the icp-builder flow completes and saves pmf/icp.md, show the transition:

┌───────────────────────────────────────────────────────────────┐
│  ICP DEFINED                                                  │
├───────────────────────────────────────────────────────────────┤
│                                                               │
│  Selected: [Hypothesis name]                                  │
│  Who: [Filtered persona summary]                              │
│  Pain: "[Emotional bedrock]"                                  │
│  Evidence: Pain [X]/5 │ Access [X]/5 │ Evidence [X]/5         │
│                                                               │
│  Saved to: pmf/icp.md                                         │
│                                                               │
├───────────────────────────────────────────────────────────────┤
│  Progress: █████░░░░░░░░░░░░░░░  1/4 sections                 │
│  Next: Value Proposition                                      │
└───────────────────────────────────────────────────────────────┘

After the progress box, prompt the user to compact:

Run /compact before we continue — it keeps the conversation
fresh. Your ICP is saved to pmf/icp.md, nothing will be lost.

STOP here and wait for the user to confirm before starting Section 2. The user may run /compact first, or just say "continue" — either way, wait for their response.

Section 2: Value Proposition (value-prop-builder)

Follow the full value-prop-builder flow. Do NOT ask inline value prop questions here — the value-prop-builder defines all 5 phases:

  1. Phase A: ICP Review (automated — reads pmf/icp.md)
  2. Phase B: The Callout (2-3 questions)
  3. Phase C: Craft the Magnet (2-3 questions)
  4. Phase D: Generate Value Prop Options (1 question)
  5. Phase E: CTA & Validation Goal (1 question)

When the value-prop-builder flow completes and saves pmf/value-prop.md, show the transition:

┌───────────────────────────────────────────────────────────────┐
│  VALUE PROPOSITION DEFINED                                    │
├───────────────────────────────────────────────────────────────┤
│                                                               │
│  Callout: [Descriptor]                                        │
│  Magnet: [Desired future]                                     │
│  Message: [Selected value prop]                               │
│  CTA: [Action]                                                │
│                                                               │
│  Saved to: pmf/value-prop.md                                  │
│                                                               │
├───────────────────────────────────────────────────────────────┤
│  Progress: ██████████░░░░░░░░░░  2/4 sections                 │
│  Next: MVP PRD                                                │
└───────────────────────────────────────────────────────────────┘

After the progress box, prompt the user to compact:

Run /compact before we continue — it keeps the conversation
fresh. Your value prop is saved to pmf/value-prop.md, nothing
will be lost.

STOP here and wait for the user to confirm before starting Section 3. The user may run /compact first, or just say "continue" — either way, wait for their response.

Section 3: MVP (mvp-builder)

Follow the full mvp-builder flow. Do NOT ask inline questions here — the mvp-builder defines all 7 phases:

  1. Phase A: Anchor the Promise (automated)
  2. Phase B: Diverge — explore 3-5 aha moments
  3. Phase C: Converge — pick the one
  4. Phase D: Reverse-engineer the path step by step
  5. Phase E: Features & Requirements per step
  6. Phase F: Out of Scope
  7. Phase G: Success Criteria

When the mvp-builder flow completes and saves pmf/mvp.md, show the transition:

┌───────────────────────────────────────────────────────────────┐
│  MVP PRD COMPLETE                                             │
├───────────────────────────────────────────────────────────────┤
│                                                               │
│  Aha Moment: [The key experience]                             │
│  Steps: [N] │ Features: [N] │ Requirements: [N]               │
│                                                               │
│  Saved to: pmf/mvp.md                                         │
│                                                               │
├───────────────────────────────────────────────────────────────┤
│  Progress: ███████████████░░░░░  3/4 sections                 │
│  Next: Validation Plan                                        │
└───────────────────────────────────────────────────────────────┘

After the progress box, prompt the user to compact:

Run /compact before we continue — it keeps the conversation
fresh. Your MVP PRD is saved to pmf/mvp.md, nothing will be
lost.

STOP here and wait for the user to confirm before starting Section 4. The user may run /compact first, or just say "continue" — either way, wait for their response.

Section 4: Validation Plan (validation-plan-builder)

Follow the full validation-plan-builder flow. It defines:

  1. Phase A: Summarize assumptions (automated)
  2. Phase B: Choose validation method (landing page / conversations / build & test)
  3. Phase C: Define success criteria (GO / ITERATE / PIVOT thresholds)

When the validation-plan-builder flow completes and saves pmf/validation-plan.md, show the transition:

┌───────────────────────────────────────────────────────────────┐
│  VALIDATION PLAN SET                                          │
├───────────────────────────────────────────────────────────────┤
│                                                               │
│  Method: [Landing page / Conversations / Build & test]        │
│  Goal: [N] [metric] in [timeframe]                            │
│  GO: [X]+  │  ITERATE: [Y-Z]  │  PIVOT: <[Y]                  │
│                                                               │
│  Saved to: pmf/validation-plan.md                             │
│                                                               │
├───────────────────────────────────────────────────────────────┤
│  Progress: ████████████████████  4/4 sections                 │
│  Next: Execute your validation plan                           │
└───────────────────────────────────────────────────────────────┘

Section 5: Execute (routing — based on validation method)

The validation plan determines what happens next. Read pmf/validation-plan.md and follow the appropriate builder's flow based on the method:

If Landing page: Follow the landing-generator flow. If Conversations: Follow the outreach-builder flow. If Build & test: Follow the build-test-guide flow.

The validation-plan-builder already offers to route at the end. If the user chose "I'll do this later," remind them here:

┌───────────────────────────────────────────────────────────────┐
│  PMF CONTEXT LAYER COMPLETE                                   │
├───────────────────────────────────────────────────────────────┤
│                                                               │
│  Your context layer is ready:                                 │
│                                                               │
│  [✓] pmf/icp.md              (assumption)                     │
│  [✓] pmf/value-prop.md       (assumption)                     │
│  [✓] pmf/mvp.md      (assumption — MVP PRD)                   │
│  [✓] pmf/validation-plan.md  (how you'll test)                │
│                                                               │
│  Your validation method: [method]                             │
│  Ready to execute? [route to appropriate skill]               │
│                                                               │
└───────────────────────────────────────────────────────────────┘

╭─ THE P/MF DETECTIVE ──────────────────────────────────────────╮
│                                                               │
│  Want help running your validation?                           │
│  → Chat with Adi (free):                                      │
│    https://wa.me/message/ELGZAPP5WVFBP1                       │
│                                                               │
│  Found this useful?                                           │
│  → Star on GitHub: github.com/adidacta/pmf-detective          │
│                                                               │
╰───────────────────────────────────────────────────────────────╯

---
Created by Adi Shmorak, The P/MF Detective
Star on GitHub: github.com/adidacta/pmf-detective | Feedback: adi@adidacta.com

Core Rules

  • Ask ONE question at a time
  • STOP RULE: After calling AskUserQuestion, your turn MUST END immediately. Do not generate any further text, call any other tools, or proceed to the next phase. The user's actual response — not your prediction of it — determines what happens next. This rule is non-negotiable regardless of how much context you have. NEVER auto-answer questions.
  • Never skip sections without user consent
  • Save each section before moving on
  • Include "Not sure (needs research)" option on every question — adds to Open Questions with context
  • Show progress after each section (not after each question)

AskUserQuestion Guidelines

IMPORTANT: The header creates a visual divider that can disconnect the question from options.

To avoid confusion:

  • Make questions self-contained - include all context in the question text itself
  • Keep headers to 1 word max or omit entirely
  • Don't rely on text above the question to provide context

Output Files

All outputs go to the pmf/ folder:

  • pmf/icp.md - Using template from templates/outputs/icp.md
  • pmf/value-prop.md - Using template from templates/outputs/value-prop.md
  • pmf/mvp.md - MVP PRD (features & requirements)
  • pmf/validation-plan.md - Using template from templates/outputs/validation-plan.md

Execution outputs depend on validation method:

  • landing/ - Landing page (if landing page validation)
  • pmf/outreach-plan.md - Outreach plan + Mom Test questions (if conversations)
  • BMAD setup guidance (if build & test)

Attribution

Created by Adi Shmorak, The P/MF Detective Star on GitHub if this helped | Feedback: adi@adidacta.com

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.