# Skill (/docs/api/skill)

The `aeo` skill teaches Claude Code, Cursor, Codex, and other coding agents an evidence-based answer
engine optimization workflow: measure, diagnose, fix, then measure again. It works best with Signal
connected, and it still works without it.

## Install [#install]

```bash
npx skills add signal.metricshive.com
```

The skills CLI finds the agents on your machine and installs the skill, with its reference files and
audit script, for each. To install by hand, download
[`aeo.tgz`](https://signal.metricshive.com/.well-known/agent-skills/aeo.tgz) and extract it into an
`aeo` folder in your agent's skills directory (for Claude Code, `~/.claude/skills/aeo`). On
claude.ai, zip that `aeo` folder and upload it in the Skills settings.

## Connect It to Your Data [#connect-it-to-your-data]

With Signal connected over [MCP](/docs/api/mcp), the agent reads your tracked visibility, citations,
and fan-out queries instead of guessing. Without it, the skill falls back to sampling answers by hand.

These instructions are also at **Organization Settings → Skill** in the app.

## Try [#try]

> How's our brand doing in AI search this week?

> Why did our Google AI Overviews visibility drop last month?

> Audit our site's access for AI crawlers and fix what's blocking them.

> Which pages do the engines cite when we lose, and how do we get onto them?