compound-product

🥈Silver

Autonomous system that reads product performance reports, identifies top priorities, and implements fixes without human intervention. Benefits operations teams by reducing manual work and accelerating issue resolution. Connects to product analytics tools and workflow automation systems.

503340Updated 2w ago

Intermediate30min to implementautomation

Saves ~480 min per use

Quick InstallView Source

git clone https://github.com/snarktank/compound-product.git

Works with:

Claude

Overview

About This Skill

https://github.com/snarktank/compound-product

How to Use

[{"step":"Connect the compound-product system to your product analytics tool (e.g., Google Analytics, Mixpanel, or internal dashboards) using the provided API endpoint or webhook. Ensure the system has read/write permissions for performance metrics and workflow automation.","tip":"Use a dedicated service account for the AI system to avoid permission conflicts. Test the API connection with a sample report first."},{"step":"Define your prioritization criteria (e.g., conversion rate drops >10%, error rates >5%, or load times >3s). Customize the [METRICS] placeholder in the prompt to match your KPIs.","tip":"Start with a narrow scope (e.g., checkout flow only) to validate the system’s accuracy before expanding to other product areas."},{"step":"Configure your workflow automation tool (e.g., Zapier, Make, or internal scripts) to allow the AI system to trigger actions like scaling services, updating configs, or rolling back deployments.","tip":"Set up approval gates for high-risk actions (e.g., database migrations) by using conditional logic in your workflow tool (e.g., 'Only proceed if error rate >15%')."},{"step":"Run the prompt template in your AI assistant (e.g., Claude or ChatGPT) with your [PLACEHOLDERS] filled in. Review the generated action plan before enabling full autonomy.","tip":"Start with a 'dry run' mode where the AI suggests fixes but doesn’t execute them. Use this to build trust and refine the system’s recommendations."},{"step":"Enable the autonomous mode and monitor the system’s actions via logs or dashboards. Adjust thresholds or criteria as needed based on real-world performance.","tip":"Set up alerts for when the AI system takes an action outside expected parameters (e.g., scaling resources beyond 50% of capacity)."}]

Use Cases

Automatically analyze daily performance reports to identify critical issues.

Generate and implement fixes for the most impactful problems without manual input.

Create pull requests for review based on prioritized tasks derived from report analysis.

Run quality checks on implemented tasks to ensure code integrity before merging.

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/snarktank/compound-product

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Act as an autonomous compound-product system. Review the latest product performance report for [PRODUCT_NAME] from [ANALYTICS_TOOL] (accessible via [API_ENDPOINT]). Identify the top 3 performance issues based on [METRICS: e.g., conversion rate, error rate, load time]. For each issue, generate a prioritized action plan with specific fixes and automation steps. Then, execute these fixes using [WORKFLOW_TOOL] (e.g., Zapier, Make, or internal scripts) without requiring manual approval. Provide a summary of actions taken and results achieved.

Example Output

### Autonomous Product Performance Fix Report
**Product:** *ShopEase Checkout Module*
**Report Date:** June 10, 2024
**Analytics Tool:** Google Analytics 4 (via API)

#### Identified Issues & Actions Taken:
1. **Issue:** Cart abandonment rate increased by 12% (from 68% to 80%) in the last 48 hours.
   - **Root Cause:** Payment gateway timeout errors (Error Code: PG-408) during peak traffic.
   - **Fix Applied:** Automatically scaled payment service pods in Kubernetes by 30% and adjusted timeout settings from 30s to 45s using Terraform scripts.
   - **Result:** Abandonment rate dropped to 71% within 2 hours. No further errors reported.

2. **Issue:** Mobile checkout page load time spiked to 4.2s (target: <3s).
   - **Root Cause:** Unoptimized image assets (avg. size: 1.8MB per image).
   - **Fix Applied:** Triggered ImageOptim via Zapier to compress all product images in the checkout flow. Updated CDN cache headers to enforce revalidation.
   - **Result:** Load time reduced to 2.8s. Performance score improved from 68 to 85 (Lighthouse).

3. **Issue:** Error rate in discount code validation jumped to 8% (from 2%).
   - **Root Cause:** Race condition in Redis cache during high-concurrency requests.
   - **Fix Applied:** Deployed a temporary Redis cluster (3 nodes) and updated cache TTL from 5m to 1m. Rolled back via ArgoCD after 1 hour.
   - **Result:** Error rate stabilized at 2.5%. No customer complaints received.

#### Summary of Actions:
- **Automated Tools Used:** Kubernetes (scaling), Terraform (config changes), Zapier (image optimization), ArgoCD (rollback).
- **Time Saved:** ~4 hours of manual debugging and implementation.
- **Impact:** 15% improvement in checkout completion rate and 22% faster load times.

**Next Steps:** Monitor metrics for 24 hours. If issues recur, escalate to the on-call team with pre-generated runbooks.