Rishikesh Ranjan — rishikeshranjan.com

// setup

Init

/ttm-init

Interview-driven onboarding that generates .taketomarket/ reference files (POSITIONING, BRAND, ICP, PRODUCT-DNA) with brand colors + logo. Supports YOLO mode for non-critical question skipping. Use when setting up takeToMarket for a new project.

manual invocation
allowed-tools:ReadWriteEditBashGlobGrepAskUserQuestion
SKILL.md ↗

Overview

ttm-init is takeToMarket's interview-driven onboarding skill. It runs a structured ~30-question interview across six sections (product/positioning, brand/voice, ICP/audience, channels, competitors, metrics/calendar), then generates the full .taketomarket/ set of reference files — POSITIONING, BRAND, ICP, CHANNELS, COMPETITORS, METRICS, CALENDAR, STATE, LEARNINGS, and PRODUCT-DNA — plus brand colors, a logo asset set, and the CLAUDE.md / AGENTS.md instruction files.

Think of it as npm init for marketing: one interactive pass produces the config that every other /ttm-* skill reads. Every downstream phase — produce, verify, ship, measure — loads these reference files as their context, so skipping init means later skills operate without your positioning invariant and drift. Run ttm-init once when setting up takeToMarket for a new project; re-run only on a real positioning shift via /ttm-positioning-shift.

It supports a YOLO mode that skips non-critical questions (using documented defaults) and a --text mode that replaces interactive prompts with plain-text numbered lists. Critical questions — positioning, manifesto beliefs/worldview, primary brand color, and logo final approval — are never skipped.

How it works

The skill reads and follows the workflow at ${CLAUDE_PLUGIN_ROOT}/workflows/setup/init.md. Its purpose is interview-driven onboarding that generates all .taketomarket/ reference files from structured questioning, orchestrating a 6-section interview with specificity validation, then generating 10 reference files (including PRODUCT-DNA) plus brand assets (colors.json + logo set) plus CLAUDE.md and AGENTS.md.

Required reading for the workflow: ${CLAUDE_PLUGIN_ROOT}/workflows/setup/init-questions.md and ${CLAUDE_PLUGIN_ROOT}/workflows/setup/init-validation.md.

The skill is backed by the helper script ${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs, which provides the first-run, config read, init, timestamp date, health, and state update subcommands the steps below call.

Step 0: First-run inline education

Read .taketomarket/CONFIG.md. Parse first_run_seen (object) and inline_education (boolean, default true).

If inline_education is false: skip this step. Else if first_run_seen.ttm-init is not true, print the explainer below verbatim, then mark this skill as seen:

node "${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs" first-run mark ttm-init

Use this exact check (bash) to decide whether to print: node "${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs" first-run check ttm-init --raw — the JSON seen field is true once the explainer has run before.

Explainer for /ttm-init

/ttm-init is the bootstrap interview. It asks you ~30 structured questions about your product, audience, positioning, and brand, then generates the .taketomarket/ reference files (POSITIONING.md, ICP.md, BRAND.md, PRODUCT-DNA.md, plus brand colors and a logo set) and writes CLAUDE.md / AGENTS.md so the runtime knows the rules. Think of it as npm init for marketing: one interactive pass produces the config that every other /ttm-* skill reads.

Why it matters now: every downstream skill — produce, verify, ship, measure — loads these reference files as their context. If you skip init, every later skill is operating without your positioning invariant and will drift. Run init once per project; re-run only on a real positioning shift via /ttm-positioning-shift.

(Canonical source: references/inline-education-blurbs.md. Embedded verbatim because workflows do not @-resolve files at runtime.)

Text-Mode Detection

Text mode (--text flag): Set TEXT_MODE=true if --text is present in $ARGUMENTS or if the runtime is not Claude Code. When TEXT_MODE is active, replace every AskUserQuestion call with a plain-text numbered list and ask the user to type their choice number.

Detection:

if echo "$ARGUMENTS" | grep -q -- '--text'; then TEXT_MODE=true; fi

If AskUserQuestion tool is not available in the current runtime, set TEXT_MODE=true.

When TEXT_MODE is active, replace each AskUserQuestion with a plain-text numbered list:

[HEADER]
[QUESTION]
  1. [OPTION_1_LABEL] -- [OPTION_1_DESCRIPTION]
  2. [OPTION_2_LABEL] -- [OPTION_2_DESCRIPTION]
  ...
Type the number of your choice:

For multiSelect questions, instruct the user: "Type the numbers of your choices separated by commas (e.g., 1,3,5):"

YOLO Mode Detection

Read .taketomarket/CONFIG.md if it exists:

node ${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs config read --raw 2>/dev/null

If yolo: true: set YOLO_MODE=true. Throughout this workflow, when a question is tagged priority: non-critical in init-questions.md, skip it and use its documented default: value.

Critical questions (positioning, manifesto beliefs/worldview, primary brand color, logo final approval) are NEVER skipped regardless of YOLO mode.

Step 1: Pre-flight

takeToMarket > INITIALIZING

Run init status check:

node ${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs init --raw

If result is "initialized":

Use AskUserQuestion:

  • header: "Existing Setup"
  • question: ".taketomarket/ already contains reference files. What would you like to do?"
  • options:
    • label: "Start fresh"
      • description: "Delete existing files and re-run the full interview"
    • label: "Update specific files"
      • description: "Choose which reference files to regenerate"
    • label: "Cancel"
      • description: "Exit without changes"
  • If "Start fresh": Delete .taketomarket/ contents (except CAMPAIGNS/ and PLAYBOOKS/) and continue to Step 2.
  • If "Update specific files": Present file picker using AskUserQuestion with multiSelect: true listing all 10 reference files (POSITIONING, BRAND, ICP, CHANNELS, COMPETITORS, METRICS, CALENDAR, STATE, LEARNINGS, PRODUCT-DNA). Run only the interview sections that map to selected files (see Question-to-Template Mapping in init-questions.md). If PRODUCT-DNA is selected, re-run the sub-workflow at ${CLAUDE_PLUGIN_ROOT}/workflows/setup/init-product-dna.md. Skip unselected sections and jump to Step 9 for generation.
  • If "Cancel": Exit with message "No changes made."

If result is "not initialized":

mkdir -p .taketomarket/CAMPAIGNS
mkdir -p .taketomarket/PLAYBOOKS

Continue to Step 2.

Step 2: Interview — Section 1: Product and Positioning

takeToMarket > SECTION 1: PRODUCT AND POSITIONING

This section collects data for POSITIONING.md.

Ask the Product Category structured question from init-questions.md Section 1 via AskUserQuestion. Then ask all 5 freeform questions from init-questions.md Section 1 in sequence.

Specificity validation for Section 1:

After collecting all answers, validate:

  • Differentiator check: FAIL if differentiator is a single adjective or contains only banned phrases. FAIL if it does not describe a specific mechanism, capability, or process.
    • PASS: "Auto-generates sprint plans from Slack conversations"
    • FAIL: "The most innovative project management tool"
  • Target audience check: FAIL if audience specifies fewer than 2 of: role, company size, industry, geography.
    • PASS: "VP Marketing at B2B SaaS companies with 50-500 employees"
    • FAIL: "Modern teams"
  • Proof points check: FAIL if fewer than 2 proof points. FAIL if any lacks a verifiable source.
    • PASS: "42% reduction in churn after implementing X (Source: Q3 2025 internal analysis)"
    • FAIL: "Customers love our product"
  • Must-not-say check: FAIL if fewer than 2 banned terms. FAIL if any lacks reasoning.
    • PASS: "'Disruptive' -- overused in our category, signals no real differentiation"
    • FAIL: Just a list of words with no reasoning

Global banned phrases — check all freeform answers against the Banned Phrases table in init-validation.md. Reject any answer containing superlatives, buzzwords, impact words, or vague qualifiers without a specific mechanism.

On validation failure, re-prompt:

Your [FIELD_NAME] is too vague. Here's what specific looks like:
  - Vague: '[FAIL_EXAMPLE]'
  - Specific: '[PASS_EXAMPLE]'
Please try again with more specificity.

Maximum 2 re-prompts per field. After 2 retries, accept with flag: <!-- SPECIFICITY_WARNING: [field] accepted after 2 retries -->.

Store all collected answers for file generation in Step 9.

Step 3: Interview — Section 2: Brand and Voice

takeToMarket > SECTION 2: BRAND AND VOICE

This section collects data for BRAND.md.

Ask the Voice Archetype and Formality Level structured questions from init-questions.md Section 2 via AskUserQuestion. Then ask all 4 freeform questions from init-questions.md Section 2 in sequence.

Specificity validation for Section 2:

  • Voice archetype check: FAIL if description is under 20 words. FAIL if no concrete personality traits.
    • PASS: "Confident but not arrogant. We explain complex concepts simply. We use data to back claims. We never talk down to the reader."
    • FAIL: "Professional and friendly"
  • Examples check: FAIL if on-brand example is under 10 words. FAIL if off-brand example is missing or identical style.
  • Banned words check: FAIL if zero banned words listed. WARN if banned words lack reasoning.

On WARN: "Your banned words could be more specific, but I'll accept it for now. You can update this later with /ttm-brand-refresh."

Step 4: Interview — Section 3: ICP and Audience

takeToMarket > SECTION 3: ICP AND AUDIENCE

This section collects data for ICP.md.

No structured questions for this section. Ask all 6 freeform questions from init-questions.md Section 3 in sequence.

Specificity validation for Section 3:

  • Segment check: FAIL if description specifies fewer than 2 of: role, company size, industry, geography.
  • Pain points check: FAIL if fewer than 2 pain points. FAIL if pain points are generic.
    • PASS: "Spends 4+ hours weekly manually reconciling campaign data across 5 analytics platforms"
    • FAIL: "Struggles with marketing"
  • JTBD check: FAIL if job description is under 10 words. FAIL if it uses banned phrases.
  • Customer language check: WARN (not FAIL) if fewer than 3 phrases provided — this data is harder to have on hand.

On WARN: "Your customer language could be more detailed, but I'll accept it for now. You can update this later with /ttm-icp-refresh."

Step 4b: PRODUCT-DNA Generation

takeToMarket > PRODUCT-DNA (WHAT + WHY)

Read and follow ${CLAUDE_PLUGIN_ROOT}/workflows/setup/init-product-dna.md.

This sub-workflow scans the codebase (WHAT) and interviews the user for manifesto + worldview (WHY), then writes .taketomarket/PRODUCT-DNA.md. It depends on POSITIONING + BRAND + ICP for context coherence and must run before brand colors + logo (which use PRODUCT-DNA's worldview for rationale generation).

Step 4c: Brand Colors

takeToMarket > BRAND COLORS

Read and follow ${CLAUDE_PLUGIN_ROOT}/workflows/setup/init-brand-colors.md.

This sub-workflow proposes 3 WCAG-checked palettes, lets the user pick, then appends colors to BRAND.md and saves .taketomarket/brand/colors.json for landing/pSEO consumption.

Step 4d: Logo

takeToMarket > LOGO

Read and follow ${CLAUDE_PLUGIN_ROOT}/workflows/setup/init-logo.md.

This sub-workflow generates SVG candidates with vision self-review (up to 3 rounds), produces a full asset set under .taketomarket/brand/, and appends a Logo section to BRAND.md.

Step 5: Interview — Section 4: Channels

takeToMarket > SECTION 4: CHANNELS

This section collects data for CHANNELS.md.

Ask the Primary Channel structured question from init-questions.md Section 4 via AskUserQuestion. Then ask all 4 freeform questions from init-questions.md Section 4 in sequence.

Specificity validation for Section 4:

  • Channel list check: FAIL if zero active channels listed. WARN if no baseline metrics provided for active channels.
  • Budget check: WARN if no budget split provided (acceptable for early-stage companies).

On WARN: "Your channel baselines could be more detailed, but I'll accept it for now. You can update this later with /ttm-health."

Step 6: Interview — Section 5: Competitors

takeToMarket > SECTION 5: COMPETITORS

This section collects data for COMPETITORS.md.

No structured questions for this section. Ask all 4 freeform questions from init-questions.md Section 5 in sequence.

Specificity validation for Section 5:

  • Competitor list check: FAIL if zero competitors named. WARN if fewer than 2 competitors (acceptable for new categories). FAIL if entries lack positioning description.
  • Positioning map check: WARN if no axes defined (can be refined later).

On WARN: "Your competitor analysis could be more detailed, but I'll accept it for now. You can update this later with /ttm-competitor-scan."

Step 7: Interview — Section 6: Metrics and Calendar

takeToMarket > SECTION 6: METRICS AND CALENDAR

This section collects data for METRICS.md and CALENDAR.md.

Ask the Attribution Model structured question from init-questions.md Section 6 via AskUserQuestion. Then ask all 7 freeform questions from init-questions.md Section 6 in sequence.

Specificity validation for Section 6:

  • Primary metric check: FAIL if metric is an output metric only (e.g., "blog posts published") without an outcome metric.
    • PASS: "Qualified pipeline generated from content marketing -- target $500K/quarter"
    • FAIL: "Number of blog posts published per month"
  • Secondary metrics check: WARN if no secondary metrics provided.
  • Calendar check: WARN if no quarterly theme provided. WARN if no cadence defined.

On WARN: "Your metrics/calendar data could be more detailed, but I'll accept it for now. You can update this later with /ttm-health."

Step 8: Confirmation Gate

takeToMarket > REVIEW

Display a summary of what was collected:

  • Product category and positioning summary (differentiator in one sentence)
  • Voice archetype and formality level
  • Primary ICP segment (role, company size, industry)
  • Active channels and primary channel
  • Top competitors named
  • Primary outcome metric and target

Use AskUserQuestion:

  • header: "Ready?"
  • question: "Here's what I'll use to generate your reference files. Ready to proceed?"
  • options:
    • label: "Generate files"
      • description: "Create all .taketomarket/ reference files"
    • label: "Revise a section"
      • description: "Go back and update specific answers"
  • If "Revise a section": Use AskUserQuestion with options listing sections 1-6. Re-run the selected section's interview questions and validation. Return to this confirmation gate.
  • If "Generate files": Continue to Step 9.

Step 9: Generate Reference Files

takeToMarket > GENERATING REFERENCE FILES

For each reference file, follow this pattern:

  1. Read the template: ${CLAUDE_PLUGIN_ROOT}/templates/reference-files/{filename}.md
  2. Generate the file content by filling ALL [GENERATED BY /ttm-init] placeholders with interview data
  3. Critical structural requirements:
    • Preserve <!-- _SUMMARY: ... --> and <!-- END_SUMMARY --> comment markers exactly as they appear in the template
    • Summary section must be under 200 words
    • Preserve all table structures from the template
    • For dates, use: node ${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs timestamp date --raw
  4. Write to .taketomarket/{FILENAME}.md

Generate files in this order:

1. POSITIONING.md (from Section 1 data)

Read template: ${CLAUDE_PLUGIN_ROOT}/templates/reference-files/positioning.md. Fill: Category, Target audience, Primary differentiator, Proof points, Must-not-say, Competitive frame, Positioning History date. Write to: .taketomarket/POSITIONING.md.

2. BRAND.md (from Section 2 data)

Read template: ${CLAUDE_PLUGIN_ROOT}/templates/reference-files/brand.md. Fill: Voice archetype, Voice attributes, Tone per context, Banned words with reasoning, Proof points, Good and bad examples. Write to: .taketomarket/BRAND.md.

3. ICP.md (from Section 3 data)

Read template: ${CLAUDE_PLUGIN_ROOT}/templates/reference-files/icp.md. Fill: Primary segment, Demographics, Psychographics, JTBD, Pain points with severity and frequency, Buying triggers, Anti-ICP, Customer language library. Write to: .taketomarket/ICP.md.

4. CHANNELS.md (from Section 4 data)

Read template: ${CLAUDE_PLUGIN_ROOT}/templates/reference-files/channels.md. Fill: Active channels with baselines, Primary channel, Dormant channels with reasons, Banned channels with reasons, Budget allocation, Channel-specific rules. Write to: .taketomarket/CHANNELS.md.

5. COMPETITORS.md (from Section 5 data)

Read template: ${CLAUDE_PLUGIN_ROOT}/templates/reference-files/competitors.md. Fill: Direct competitors (positioning, strength, weakness), Positioning map axes and positions, Share of voice baseline. Write to: .taketomarket/COMPETITORS.md.

6. METRICS.md (from Section 6 data -- metrics portion)

Read template: ${CLAUDE_PLUGIN_ROOT}/templates/reference-files/metrics.md. Fill: Primary outcome metric (metric/target/window/source), Secondary metrics, Leading indicators, Baselines, Attribution model. Write to: .taketomarket/METRICS.md.

7. CALENDAR.md (from Section 6 data -- calendar portion)

Read template: ${CLAUDE_PLUGIN_ROOT}/templates/reference-files/calendar.md. Fill: Quarterly themes, Launch calendar, Always-on cadence, Blackout dates. Write to: .taketomarket/CALENDAR.md.

8. STATE.md (template copy with timestamp)

Read template: ${CLAUDE_PLUGIN_ROOT}/templates/reference-files/state.md. Special handling: Copy template and update ONLY the last_updated frontmatter field:

node ${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs timestamp date --raw

Do NOT fill interview data into STATE.md -- it is initialized with default values. Write to: .taketomarket/STATE.md.

9. LEARNINGS.md (template copy with zeroed counters)

Read template: ${CLAUDE_PLUGIN_ROOT}/templates/reference-files/learnings.md. Special handling: Copy template verbatim. Set Summary fields to:

  • Total lessons: 0
  • Last lesson date: none
  • Top pattern: none

Do NOT fill interview data -- learnings are populated during campaign Learn phases. Write to: .taketomarket/LEARNINGS.md.

Step 10: Generate Instruction Files

takeToMarket > GENERATING INSTRUCTION FILES

Read ${CLAUDE_PLUGIN_ROOT}/templates/claude-md.md and write to CLAUDE.md in the project root. Read ${CLAUDE_PLUGIN_ROOT}/templates/agents-md.md and write to AGENTS.md in the project root.

These are static copies -- do NOT customize with interview data. The positioning-as-invariant enforcement rules are already baked into the templates. Runtime context loading handles dynamic positioning data via POSITIONING.md.

Step 11: Post-Init Validation

takeToMarket > VALIDATING

Run health check:

node ${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs health

Parse the JSON result. Verify all checks have status "pass".

If any check has status "fail" or "missing":

  • Report which files failed
  • Attempt to regenerate only the failed files by re-reading the template and re-writing
  • Re-run health check

Once all checks pass, update state:

node ${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs state update status initialized

Warning tracking: If any interview section accumulated 3 or more SPECIFICITY_WARNING flags during the interview, append to .taketomarket/STATE.md under a ## Follow-up Needed heading:

  • Section [N] ([name]): [count] specificity warnings -- recommend re-running with [relevant /ttm-* command]

Use these commands per section:

  • Section 1 (Positioning): /ttm-positioning-check
  • Section 2 (Brand): /ttm-brand-refresh
  • Section 3 (ICP): /ttm-icp-refresh
  • Section 4 (Channels): /ttm-health
  • Section 5 (Competitors): /ttm-competitor-scan
  • Section 6 (Metrics/Calendar): /ttm-health

Step 12: Summary

takeToMarket > INITIALIZED

Reference files created:
  .taketomarket/POSITIONING.md
  .taketomarket/BRAND.md
  .taketomarket/ICP.md
  .taketomarket/CHANNELS.md
  .taketomarket/STATE.md
  .taketomarket/METRICS.md
  .taketomarket/COMPETITORS.md
  .taketomarket/CALENDAR.md
  .taketomarket/LEARNINGS.md
  .taketomarket/PRODUCT-DNA.md

Brand assets created:
  .taketomarket/brand/colors.json
  .taketomarket/brand/logo.svg
  .taketomarket/brand/logo-mark.svg
  .taketomarket/brand/logo-wordmark.svg
  .taketomarket/brand/logo-mono-light.svg
  .taketomarket/brand/logo-mono-dark.svg
  .taketomarket/brand/logo-favicon.svg
  .taketomarket/brand/logo@1x.png
  .taketomarket/brand/logo@2x.png

Instruction files created:
  CLAUDE.md
  AGENTS.md

All health checks passed. Your marketing operating system is ready.

Next step: Run /ttm-new-campaign to create your first campaign.

Success criteria

  • Pre-flight check completed (existing init detected or directories created)
  • All 6 interview sections completed with specificity validation
  • PRODUCT-DNA sub-workflow run; .taketomarket/PRODUCT-DNA.md written
  • Brand-colors sub-workflow run; BRAND.md ## Colors section filled + .taketomarket/brand/colors.json written
  • Logo sub-workflow run; .taketomarket/brand/ asset set produced + BRAND.md ## Logo section filled
  • Confirmation gate passed
  • 10 reference files generated in .taketomarket/ with _SUMMARY/END_SUMMARY preserved
  • CLAUDE.md and AGENTS.md copied to project root
  • Health check passes (all 12 checks: taketomarket_dir + campaigns_dir + 10 files)
  • STATE.md status set to "initialized"

Output

  • .taketomarket/POSITIONING.md
  • .taketomarket/BRAND.md
  • .taketomarket/ICP.md
  • .taketomarket/CHANNELS.md
  • .taketomarket/STATE.md
  • .taketomarket/METRICS.md
  • .taketomarket/COMPETITORS.md
  • .taketomarket/CALENDAR.md
  • .taketomarket/LEARNINGS.md
  • .taketomarket/PRODUCT-DNA.md
  • .taketomarket/brand/colors.json
  • .taketomarket/brand/logo.svg (plus mark/wordmark/mono/favicon variants and PNG raster)
  • CLAUDE.md
  • AGENTS.md

What if this doesn't fit?

Looks like /ttm-init can't do that yet.

  • Want a new skill? /ttm-request-skill
  • Existing skill needs work? /ttm-improve-skill