Audience Skill Runner

Run a single audience skill against a target URL.

Arguments: <skill-name> <url>

Parse $ARGUMENTS to extract the skill name (first word) and URL (second word).

Server: http://localhost:3500

Instructions

Read the skill definition from packs/audience/skills/<skill-name>.md to understand what this skill does and what events to emit.

Then execute that single skill following the same process as the full pipeline:

1. POST a "running" status event: curl -s -X POST http://localhost:3500/api/packs/audience/events -H 'Content-Type: application/json' -d '{"skill":"<skill-name>","component":"generic","props":{"title":"Running <skill-name>...","data":"Working on it"},"status":"running"}'

2. Do the research using WebFetch (for crawling the URL), WebSearch (for market/audience research), and Stripe API if available (source .env.local 2>/dev/null; curl -s "https://api.stripe.com/v1/customers?limit=100" -u "$STRIPE_API_KEY:")

3. DELETE the running event: curl -s -X DELETE "http://localhost:3500/api/packs/audience/events?skill=<skill-name>"

4. POST completed events with real data to http://localhost:3500/api/packs/audience/events

Refer to the full pipeline command at .claude/commands/audience.md for the detailed steps for each skill (persona-builder, segment-analysis, customer-research, free-tool-strategy).

Use real data from WebFetch and WebSearch. Never hardcode data. POST events progressively so the UI updates in real-time.