AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Growth Product Context

skill-meirdick-growth-product-context-growth-product-context · by meirdick

>-

No reviews yet
0 installs
26 views
0.0% view→install

Install

$ agentstack add skill-meirdick-growth-product-context-growth-product-context

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-meirdick-growth-product-context-growth-product-context)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
6mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Growth Product Context? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Product Context — Journey Discovery Engine

> Free skill from Growth Engineering. For auto-instrumentation, auditing, and MCP tools, install the full package.

Context

This skill scans a Laravel application's routes, controllers, and Eloquent models to discover user journeys and propose a product context with event names following the AARRR (Pirate Metrics) framework.

Scope: Laravel applications using Eloquent, Blade, Inertia, or Livewire. Outputs .growth/PRODUCT.md and per-domain journey files.

Activate when:

  • The user asks to "map journeys", "define product context", or "what should I track"
  • .growth/PRODUCT.md exists but still has placeholder values (e.g., [Your Product Name])
  • The user wants to discover or define their product's user journeys

Prerequisites:

  1. A Laravel application with routes, controllers, and models
  2. Laravel Boost installed (provides list-routes, database-schema, and tinker tools)

If .growth/PRODUCT.md doesn't exist yet, create it during Step 12.


Rules

Key Principles

  1. Propose, don't ask — Present concrete proposals based on what the codebase reveals. Never show blank templates.
  2. Confirm per domain — Present each journey domain separately and get confirmation before moving on.
  3. Incremental refinement — Start with what the code reveals, then let the user correct and refine.
  4. Plan before writing — Steps 1–10 are read-only discovery. Enter plan mode at the start and do not write any files until the user approves the final plan.

Step 1: Enter Plan Mode

Enter plan mode. All discovery happens before any files are written.

Step 2: Discover Routes

Read routes/web.php and routes/api.php. Check for Folio page-based routing in resources/views/pages/. Use the list-routes tool from Laravel Boost to get the full route table. Present a summary to the user.

Step 3: Map Controllers

For each route group, read the associated controllers. Note which models are referenced, events fired, jobs dispatched, notifications sent, and form request classes used.

Step 4: Identify Domain Entities

List all models in app/Models/, read their relationships/casts/scopes, and use database-schema to inspect table structures. Identify the central model (most relationships, most user interaction). Ask the user to confirm.

Step 5: Detect Frontend Stack

Check composer.json for Inertia or Livewire, package.json for frontend framework, and resources/views/ for Blade views. Report the detected stack.

Step 6: Classify Routes into Journey Domains

Classify routes using the AARRR framework: Acquisition, Activation, Retention, Revenue, Referral.

Read references/classification-rules.md for the route-to-domain table and the Acquisition vs Activation boundary rules.

Present classified routes grouped by domain. Ask the user to confirm or reassign any misclassified routes.

Step 7: Propose Journeys

For each confirmed domain, propose a journey flow with specific event names derived from the controllers.

Read references/examples.md for journey flow format and examples.

Present each journey and ask: "Does this flow match how users actually use the product?", "Are there any steps I'm missing?", "Should any events be renamed?"

Step 8: Identify Core Action

Based on the central model and the value delivery journey, propose the core action. Ask the user to confirm or provide an alternative.

Step 9: Detect Usage Pattern

Examine the codebase for signals that reveal how frequently users interact with the product.

Read references/classification-rules.md for the usage pattern detection table and success metrics per pattern.

Present the detected pattern with evidence. Ask the user to confirm or correct.

Step 10: Define Success Metric

Based on the core action AND the detected usage pattern, propose a success metric. Do not default to Weekly Active Users — choose the metric that fits the pattern.

Read references/classification-rules.md for pattern-to-metric mapping.

Step 11: Present Consolidated Plan

Present a single consolidated plan summarizing everything discovered and confirmed:

  • Product overview (name, one-liner, stage, value prop)
  • Central model and core action
  • Usage pattern and success metric
  • All journey domains with their flows and event names

Ask the user to approve the plan before proceeding. Do not write any files until the user explicitly approves.

Step 12: Write Results

After the user approves the plan, exit plan mode and write the files:

  1. Create .growth/ directory if it doesn't exist, with subdirectories: journeys/, experiments/, research/, decisions/, channels/, segments/, digests/
  2. Write .growth/PRODUCT.md with confirmed values
  3. Create journey files in .growth/journeys/ — one file per domain

Read references/output-formats.md for the PRODUCT.md and journey file templates.

Read references/examples.md for complete output examples.


Gotchas

  • Folio (page-based) routes won't show up in list-routes. If the project uses Laravel Folio, routes live in resources/views/pages/ as Blade files, not in routes/web.php. Scan that directory separately or you'll miss half the user journeys.
  • API routes may be the primary user interface. Don't assume routes/web.php is where the action is. SPAs with Inertia or mobile apps often have all meaningful user interactions in routes/api.php. Check both.
  • The "central model" isn't always User. In a CRM it's Lead or Contact. In an e-commerce app it's Order. In a project management tool it's Project or Task. Look at relationship density and controller usage, not just the name.
  • Not every route is a trackable journey step. Admin routes, health checks, webhook receivers, and API resource endpoints for internal tooling shouldn't be classified into AARRR stages. Filter these out early.
  • Laravel Boost tools must be available. This skill depends on list-routes, database-schema, and tinker from Laravel Boost. If these tools aren't available, the discovery phase will be incomplete. Check for Boost before starting.

Anti-patterns

  • Don't show blank templates. Never present an empty PRODUCT.md and ask the user to fill it in. Always propose concrete values based on what the codebase reveals.
  • Don't ask per-field questions. Scan the code and propose answers. Let the user correct, not create.
  • Don't classify without reading controllers. Route URLs alone are insufficient. Read the controller code to understand what each endpoint actually does.
  • Don't default to WAU. Match the success metric to the detected usage pattern.
  • Don't skip confirmation. Present each domain's journey separately and get explicit confirmation before moving on.
  • Don't invent events. Every proposed event name should map to a real controller action or route discovered in the codebase.

References

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.