Rishikesh Ranjan — rishikeshranjan.com

// 9-phase lifecycle

Review

/ttm-review

Review phase: present assets with structured review checklist for human evaluation. Use after verification to get human approval.

phase 4manual invocation
allowed-tools:ReadWriteBashGlobGrepAskUserQuestion
SKILL.md ↗

Overview

The ttm-review skill runs the Review phase of the takeToMarket lifecycle. It presents verified campaign assets to a human with a structured review checklist, surfacing each asset's content preview alongside its verification gate results so a person can sanity-check tone, claims, and channel fit before anything ships. Invoke it after verification with ttm-review [campaign-slug].

Review is human-in-the-loop sign-off. Automated gates catch invariant violations but not the subjective call about whether an asset is the right one for this moment. So instead of auto-shipping, the system stops, surfaces context (diffs plus gate results), and waits for your decision. For each asset you can approve it, send it back with structured revision feedback, or reject it outright. The hero asset is reviewed first in full detail; derivatives follow in a batch.

When an asset needs changes, review writes structured feedback files and updates the campaign manifest with a per-asset review_status, then directs you to run the fix command on the revised assets. Note that .taketomarket/POSITIONING.md is treated as read-only throughout this workflow.

How it works

The skill reads and follows the workflow at ${CLAUDE_PLUGIN_ROOT}/workflows/lifecycle/review.md. Its full content is reproduced below.

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-review 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-review

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

Explainer for ttm-review

ttm-review is human-in-the-loop sign-off. After verify passes the automated wall, review surfaces the asset with diff context and the gate results so you can sanity-check tone, claims, and channel fit before it ships. You can approve, reject with feedback, or escalate to a positioning shift.

Why it matters: automated gates catch invariant violations but not subjective calls about whether an asset is the right one for this moment. Review is the human commit-message step: the system stops, surfaces context, and waits for your decision rather than auto-shipping.

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

Purpose

Review workflow for ttm-review. Presents verified assets with structured review checklist for human evaluation (LIFE-10). Collects per-asset outcomes (Approve/Revise/Reject per D-03) with hero-first ordering (D-02). Structured revision feedback on Revise (D-12). Auto-triggers ttm-fix for revised assets (D-15).

Required reading

  • ${CLAUDE_PLUGIN_ROOT}/references/context-loading.md
  • ${CLAUDE_PLUGIN_ROOT}/references/review-checklist.md

Constraints: POSITIONING.md is READ-ONLY

POSITIONING.md is an architectural invariant. If you detect positioning drift:

  • In verify: use the Escalate option to launch ttm-positioning-shift
  • In other workflows: flag the issue and recommend running ttm-positioning-check

Only ttm-positioning-shift and ttm-init may modify POSITIONING.md.

Process

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.

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:

Step 1: Load Context

takeToMarket > LOADING CONTEXT

Extract SLUG from $ARGUMENTS (strip --text flag if present):

SLUG=$(echo "$ARGUMENTS" | sed 's/--text//g' | xargs)

If SLUG is empty, error: "Usage: /ttm-review [campaign-slug]. Provide a campaign slug." Exit.

Load Tier 1 summaries from all 9 reference files (lines 1 to <!-- END_SUMMARY -->):

  • .taketomarket/POSITIONING.md
  • .taketomarket/BRAND.md
  • .taketomarket/ICP.md
  • .taketomarket/CHANNELS.md
  • .taketomarket/STATE.md (frontmatter only)
  • .taketomarket/CALENDAR.md
  • .taketomarket/COMPETITORS.md
  • .taketomarket/METRICS.md
  • .taketomarket/LEARNINGS.md

Tier 2: None for review (per context-loading.md matrix -- review is human-driven).

Load campaign-specific files (always full-load per context-loading.md rule 4):

  • .taketomarket/CAMPAIGNS/${SLUG}/STATE.md
  • .taketomarket/CAMPAIGNS/${SLUG}/BRIEF.md

Load MANIFEST.json:

MANIFEST_PATH=".taketomarket/CAMPAIGNS/${SLUG}/MANIFEST.json"

Read .taketomarket/CAMPAIGNS/${SLUG}/MANIFEST.json. If the file does not exist, error: "No production manifest found for campaign '${SLUG}'. Run /ttm-produce first." Exit.

Load VERIFICATION.md:

Read .taketomarket/CAMPAIGNS/${SLUG}/VERIFICATION.md. If the file does not exist, error: "No verification report found for campaign '${SLUG}'. Run /ttm-verify first." Exit.

Step 2: Validate Campaign State

takeToMarket > VALIDATING CAMPAIGN

Check campaign exists:

node "${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs" campaign state "${SLUG}" --raw

If result shows exists: false: Tell the user the campaign does not exist and suggest running /ttm-new-campaign first. Exit.

Read campaign state. Check the phase field:

  • If phase is NOT "verified": Warn the user: "Campaign is in phase '${PHASE}'. Expected 'verified' before review. Running /ttm-review now may produce incomplete results. Proceed?" Wait for user confirmation via AskUserQuestion (or text-mode prompt). If user declines, exit.

Determine review run number:

Read review.run_count from state. If null or 0, set RUN_NUMBER=1. Otherwise set RUN_NUMBER to review.run_count + 1.

Step 3: Load Assets and Verification Results

takeToMarket > LOADING ASSETS AND GATE RESULTS

Parse MANIFEST.json for the asset list. The manifest contains a hero object and a derivatives array:

HERO = { name: manifest.hero.name, file: manifest.hero.file, type: manifest.hero.type }
DERIVATIVES = []
for each derivative in manifest.derivatives:
    DERIVATIVES.push({ name: derivative.name, file: derivative.file, type: derivative.type })
ALL_ASSETS = [HERO] + DERIVATIVES

Parse VERIFICATION.md for gate results. Extract the summary table to get per-asset PASS/WARN/FAIL counts and overall result for each gate. Build a lookup:

GATE_RESULTS[asset_name] = {
  gates: [ { gate_name, gate_id, result: PASS|WARN|FAIL }, ... ],
  pass_count, warn_count, fail_count,
  overall: PASS|WARN|FAIL
}

Load content previews from disk. For each asset:

  • Read the file from .taketomarket/CAMPAIGNS/${SLUG}/${asset.file}
  • Extract the first ~500 characters for the hero, first ~300 characters for derivatives (D-04)
  • If an asset file is missing from disk, display warning and skip it

If ALL asset files are missing: Error: "No asset files found on disk. Re-run /ttm-produce." Exit.

Step 4: Review Hero Asset (D-02)

takeToMarket > REVIEW: Hero Asset

Display the hero asset in full detail:

## [HERO_ASSET_NAME]
File: .taketomarket/CAMPAIGNS/${SLUG}/${HERO_FILE}

### Gate Summary
| Gate | Result |
|------|--------|
| Positioning Drift (GATE-01) | [PASS|WARN|FAIL] |
| Claim Accuracy (GATE-02) | [PASS|WARN|FAIL] |
| Voice Drift (GATE-03) | [PASS|WARN|FAIL] |
| Outcome Alignment (GATE-04) | [PASS|WARN|FAIL] |
| Funnel Integrity (GATE-05) | [PASS|WARN|FAIL|N/A] |
| UTM Hygiene (GATE-06) | [PASS|WARN|FAIL|N/A] |
| Compliance (GATE-07) | [PASS|WARN|FAIL|N/A] |
| Competitor Collision (GATE-08) | [PASS|WARN|FAIL] |
| ICP Fit (GATE-09) | [PASS|WARN|FAIL] |
| Format Correctness (GATE-10) | [PASS|WARN|FAIL] |

### Content Preview
[First ~500 characters of the asset]
[Full file: .taketomarket/CAMPAIGNS/${SLUG}/${HERO_FILE}]

Present the 4 mandatory review questions from review-checklist.md. Use AskUserQuestion (or text-mode prompt) for each question. All 4 are REQUIRED -- the reviewer must answer each one before selecting an outcome.

Question 1: Positioning Reinforcement

Using AskUserQuestion (or text-mode), ask: "Does this asset reinforce your positioning -- your primary differentiator, target audience, and category? Identify any claims that go beyond or contradict POSITIONING.md." Record the answer.

Question 2: Outcome Realism

"Is the outcome metric target realistic given this content? Could this asset plausibly drive the defined outcome (e.g., demo requests, signups, pipeline)?" Record the answer.

Question 3: Claim Substantiation

"Are all factual claims, statistics, and proof points substantiated by approved data in BRAND.md? Flag any unverifiable or unsourced claims." Record the answer.

Question 4: Competitor Differentiation

"Does this asset clearly differentiate from competitors? Could a competitor publish this same content with their name swapped in?" Record the answer.

Freeform Notes (Optional)

"Any additional feedback, concerns, or suggestions? (Press enter to skip)" Record notes if provided.

Collect Hero Outcome (D-03)

Present outcome selection using AskUserQuestion (or text-mode numbered list):

Select outcome for [HERO_ASSET_NAME]:
1. Approve -- asset is ready for shipping
2. Revise -- asset needs changes (you will provide structured feedback next)
3. Reject -- asset is fundamentally wrong (final, no fix loop)

If outcome is Approve: Set hero review_status = "approved" in memory.

If outcome is Revise: Collect structured revision feedback per D-12 (see review-checklist.md):

  1. Ask: "Which review questions did this asset fail? (select numbers, e.g., 1,3)" Options: 1=Positioning reinforcement, 2=Outcome realism, 3=Claim substantiation, 4=Competitor differentiation
  2. Ask: "Severity? 1=Minor (small adjustments), 2=Major (significant rewrite), 3=Critical (fundamental approach is wrong)"
  3. Ask: "What specifically needs to change? Be as concrete as possible -- reference specific sections, sentences, or claims."

Set hero review_status = "needs-fix" in memory.

Write the structured feedback to .taketomarket/CAMPAIGNS/${SLUG}/REVIEW-FEEDBACK-${HERO_NAME}.md:

# Review Feedback: [HERO_ASSET_NAME]

**Campaign:** [SLUG]
**Date:** [ISO_TIMESTAMP]
**Reviewer outcome:** Revise
**Severity:** [Minor|Major|Critical]

## Failed Checklist Items
- [x] Positioning reinforcement (if selected)
- [x] Outcome realism (if selected)
- [x] Claim substantiation (if selected)
- [x] Competitor differentiation (if selected)

## Specific Feedback
[User's freeform feedback text]

## Gate Failures (from VERIFICATION.md)
[Extract relevant WARN/FAIL findings for this asset from VERIFICATION.md]

If outcome is Reject: Ask: "Why are you rejecting this asset? (This will be logged permanently.)" Set hero review_status = "rejected" in memory. Record the rejection reason.

Step 5: Batch Review Derivatives (D-02, D-13)

If there are no derivative assets, skip this step.

takeToMarket > REVIEW: Derivative Assets

Reviewing ${N} derivatives. Hero asset for reference: ${HERO_FILE}

For each derivative asset:

Display Abbreviated View

Show abbreviated gate summary (PASS count / WARN count / FAIL count -- not full table):

## [DERIVATIVE_NAME]
File: .taketomarket/CAMPAIGNS/${SLUG}/${DERIVATIVE_FILE}
Gates: ${PASS_COUNT} PASS / ${WARN_COUNT} WARN / ${FAIL_COUNT} FAIL

### Content Preview
[First ~300 characters of the asset]
[Full file: .taketomarket/CAMPAIGNS/${SLUG}/${DERIVATIVE_FILE}]

Review Questions (Batch Mode)

Present the same 4 mandatory questions from review-checklist.md. In batch mode, answers can be shorter but all 4 are still required.

Freeform Notes (Optional)

"Any additional feedback? (Press enter to skip)"

Collect Derivative Outcome

Same 3 outcome options (Approve/Revise/Reject) per D-03.

  • If Approve: Set derivative review_status = "approved" in memory.
  • If Revise: Collect same structured revision feedback (failed items, severity, specific feedback). Write to .taketomarket/CAMPAIGNS/${SLUG}/REVIEW-FEEDBACK-${DERIVATIVE_NAME}.md. Set derivative review_status = "needs-fix" in memory.
  • If Reject: Collect rejection reason. Set derivative review_status = "rejected" in memory.

Step 6: Update MANIFEST.json with Review Status (D-11, D-13)

Read MANIFEST.json from .taketomarket/CAMPAIGNS/${SLUG}/MANIFEST.json.

Add review_status field to hero and each derivative:

  • "approved" for Approve outcomes
  • "needs-fix" for Revise outcomes
  • "rejected" for Reject outcomes

For Revise outcomes, also add review_feedback_file pointing to the feedback file:

{
  "hero": {
    ...existing fields...,
    "review_status": "approved|needs-fix|rejected",
    "review_feedback_file": "REVIEW-FEEDBACK-[NAME].md"
  },
  "derivatives": [
    {
      ...existing fields...,
      "review_status": "approved|needs-fix|rejected",
      "review_feedback_file": "REVIEW-FEEDBACK-[NAME].md"
    }
  ]
}

Write the updated MANIFEST.json back to disk.

Step 7: Update Campaign State

Determine overall result:

  • approved -- all assets approved (no Revise, no Reject)
  • needs-fix -- all non-rejected assets need fix
  • mixed -- some approved, some need fix, some rejected
TIMESTAMP=$(node "${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs" timestamp --raw)
node "${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs" campaign update "${SLUG}" review.run_count "${RUN_NUMBER}"
node "${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs" campaign update "${SLUG}" review.last_run "$TIMESTAMP"
node "${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs" campaign update "${SLUG}" review.overall_result "[approved|mixed|needs-fix]"
node "${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs" campaign update "${SLUG}" phase reviewed
node "${CLAUDE_PLUGIN_ROOT}/bin/ttm-tools.cjs" campaign update "${SLUG}" phase.reviewed "$TIMESTAMP"

Step 8: Display Completion and Next Steps (D-15)

takeToMarket > REVIEW COMPLETE

Run: ${RUN_NUMBER} | Date: ${ISO_DATE}
Assets reviewed: ${TOTAL}
  Approved: ${APPROVE_COUNT}
  Needs fix: ${REVISE_COUNT}
  Rejected: ${REJECT_COUNT}

If any assets were rejected:

Rejected assets logged. These will not enter the fix loop.

If all assets approved:

All assets approved! Run `/ttm-ship ${SLUG}` to launch.

If any assets need fix (D-15):

Run `/ttm-fix ${SLUG}` to fix ${REVISE_COUNT} asset(s) with review feedback.

Per D-15, the auto-trigger is an instruction to the user to run /ttm-fix next. The review workflow completes here -- /ttm-fix is a separate command invoked by the user (review.md is NOT forked and cannot directly invoke /ttm-fix).

If mixed results (some approved, some need fix):

${APPROVE_COUNT} asset(s) approved and ready for shipping.
${REVISE_COUNT} asset(s) need fixes. Run `/ttm-fix ${SLUG}` to address review feedback.
After fixes, approved + fixed assets can ship together via `/ttm-ship ${SLUG}`.

Success criteria

  • All assets from MANIFEST.json presented with gate summary and content preview
  • 4 mandatory review questions asked per asset (positioning, outcome, claims, differentiation)
  • Hero reviewed first in full detail, then derivatives in batch (D-02)
  • Per-asset outcome collected: Approve, Revise, or Reject (D-03, D-13)
  • Structured revision feedback captured for Revise outcomes (D-12)
  • REVIEW-FEEDBACK-[NAME].md written for each revised asset
  • MANIFEST.json updated with review_status per asset (D-11)
  • Campaign state updated: review.run_count, review.last_run, review.overall_result
  • Campaign phase advanced to "reviewed"
  • User directed to /ttm-fix for any revised assets (D-15)

Output

  • .taketomarket/CAMPAIGNS/${SLUG}/MANIFEST.json (updated with review_status per asset)
  • .taketomarket/CAMPAIGNS/${SLUG}/REVIEW-FEEDBACK-*.md (one per revised asset)

Step: Landing/pSEO quality gates (conditional)

If any asset under review is a landing-page or pSEO asset (detect by path under landing_path config):

Read and follow ${CLAUDE_PLUGIN_ROOT}/workflows/site/quality-gates.md. Pass each gated asset through gates 1-6.

If any gate fails: report failure to user with gate output. Recommend /ttm-fix.

What if this doesn't fit?

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

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