monitor-experiments

Monitor Amplitude experiments and triage them by importance.

142|16|Updated Mar 20, 2026
One-click install
npx skills add https://github.com/amplitude/builder-skills --skill monitor-experiments
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: monitor-experiments
Source: https://github.com/amplitude/builder-skills/tree/main/analytics-skills/skills/monitor-experiments
Command: npx skills add https://github.com/amplitude/builder-skills --skill monitor-experiments

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill automates the routine, error-prone work of checking experiment health across Amplitude projects, triaging which tests need attention, and producing a concise, non-technical report with clear next steps so stakeholders can act quickly.

Core Features & Use Cases

  • Portfolio scanning & filtering: Search active and recently completed experiments, exclude drafts/stale tests, and prioritize by recency and impact.
  • Metadata & metric resolution: Batch-fetch experiment metadata, resolve metric IDs to human-readable names, and extract primary metric summaries with validity flags.
  • Triage, reporting & deep-dive: Produce a top-of-report summary table with verdicts and action items, deep-dive into up to three experiments with data quality checks, and list needs-setup or configuration issues.
  • Use case: Run a weekly experiment health report for a product team to surface Ship/Iterate/Abandon recommendations, flag SRM or config errors, and identify experiments missing metrics.

Quick Start

Ask the experiment monitor to check the specified Amplitude project for running and recently completed experiments and return a prioritized summary table with action items and deep-dives.

Frequently Asked Questions about monitor-experiments

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I monitor active Amplitude experiments and triage which tests need attention?

To monitor Amplitude experiments, you can scan active and recently completed tests across projects, batch-fetch metadata, resolve metric IDs, and prioritize tests by recency and impact to triage which ones need attention.

What is the best way to generate a weekly experiment health report for A/B testing?

Generating a weekly experiment health report involves scanning your A/B testing portfolio, extracting primary metric summaries with validity flags, and producing a summary table with Ship, Iterate, or Abandon verdicts and action items.

How do I check data quality and validity flags for running experiments?

Checking data quality for running experiments involves batch-querying primary metric summaries and validity flags to detect configuration errors, identify missing metrics, and surface SRM or other data quality issues requiring setup fixes.

Can I search experiments across multiple Amplitude projects and filter out drafts or stale tests?

Yes, you can search experiments across Amplitude projects and filter results by excluding draft and stale tests, allowing you to prioritize your portfolio by recency and impact for focused reporting and deep dives.

How does experiment triage resolve metric IDs and summarize primary metrics?

Experiment triage resolves metric IDs to human-readable names by batch-fetching experiment metadata and extracting primary metric summaries with validity flags to produce concise, non-technical reports with clear next steps.

What should I do when an A/B test has configuration errors or missing metrics?

When an A/B test has configuration errors or missing metrics, you should use automated triage to list needs-setup issues, flag SRM errors, and output prioritized action items so stakeholders can apply configuration fixes quickly.