Skip to content
SkillModel-led1,115 tokens in use

Closed-Loop Analytics Upgrade

Check whether a marketing change worked by pulling analytics after the change window, comparing baseline to candidate, and patching the playbook from the evidence.

3,862 stars on the repository MIT Licenseericosiu/ai-marketing-skills · closed-loop-analytics-upgrade
npx skills add ericosiu/ai-marketing-skills -s closed-loop-analytics-upgrade
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Closed-Loop Analytics Upgrade costs

25tokens idle
1,115tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter8/15
  • Structure19/25
  • Examples9/15
  • Output15/15
  • Size15/15
  • Resources0/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity88
  • Maintenance44
  • Community56
  • Maturity28
  • Momentum40

Closed-Loop Analytics Upgrade holds that a workflow is not a closed loop until it checks whether the change worked. After a skill, prompt or campaign change ships, it pulls platform analytics for a baseline window and a candidate window, compares them, and proposes a patch to the playbook only when the data supports it.

The skill lists what to track on four surfaces: X, YouTube, SEO and answer-engine visibility, and revenue and outbound. It requires a fixed set of readback fields for every promoted change and ends each run with a verdict of promote, keep testing, rollback or unproven. Read-only pulls are fine; any external write waits for approval.

What You Get

  • A readback report in a fixed Markdown template with verdict, change tested, data pulled, baseline versus candidate table, caveats, patch and next readback date
  • A metric checklist for X, YouTube, SEO and AEO, and revenue and outbound
  • Required readback fields such as owner, windows, primary metric, winner and confounders
  • Clear promotion and no-promotion rules
  • A safety list of external writes that need approval

Common Use Cases

  • Judge whether a new post hook formula raised impressions and profile clicks after two weeks
  • Decide whether a YouTube title and thumbnail change improved click-through and retention
  • Check a content refresh against Search Console clicks and position before keeping it
  • Review an outbound sequence patch against positive replies and booked meetings

What Makes It Unique

Many analytics skills report numbers. This one governs the decision: it promotes a change only when it beats baseline on the primary metric or exposes a repeatable signal without worse downside metrics, and it refuses promotion for low volume, dirty attribution, seasonality, connector failure or when only the author liked the result. Every verdict comes with a next readback date.

Under The Hood

It is a prompt-only skill with no scripts. The agent reads analytics from whatever sources you have connected, such as Search Console, GA4, Ahrefs, HubSpot, Gong, Instantly, Metricool or social platforms, and writes a Markdown readback. The skill allows read-only pulls and says that posting, publishing, ad changes, CRM or outbound changes, emails, DMs and credential changes all need your approval first.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers