Customization
Before executing, check for user customizations at:
~/.Codex/skills/PAI/USER/SKILLCUSTOMIZATIONS/SearchIntentAutomation/
If this directory exists, load and apply any PREFERENCES.md, configurations, or resources found there. These override default behavior. If the directory does not exist, proceed with skill defaults.
SearchIntentAutomation
Turn a seed topic into an actionable research and automation workflow.
Workflow Routing
| Trigger | Workflow |
|---|---|
| Extract a reusable workflow from a reel or creator demo | Workflows/ExtractFromSource.md |
| Build a keyword, question, and automation stack for a topic or offer | Workflows/BuildOpportunityMap.md |
| Run the workflow end-to-end with CLI coding agents | Workflows/RunWithCliAgents.md |
Quick Reference
- Step 1: Use
Ubersuggestto map demand, keyword variants, SEO opportunities, and competitor gaps. - Step 2: Use
AnswerThePublicto map real audience questions, comparisons, and prepositions around the same seed term. - Step 3: Use
Playwright MCPto capture browser-only research data and a Python CLI pipeline to normalize, merge, and route outputs. - Step 3a: Hand off browser results with the capture manifest contract in
Tools/CaptureStatusContract.md. - Step 4: Optional AI layer: generate content briefs or landing-page drafts from the validated keywords/questions.
- If the pipeline gets stuck, stop and present a taxonomy-based checkpoint with:
- Recommended option 1
- Recommended option 2
- Custom user direction
Loaded Context
- Reel evidence and extraction trace:
ReelEvidence.md - Core workflow details:
Workflows/BuildOpportunityMap.md - Non-interactive agent execution:
Workflows/RunWithCliAgents.md - Failure and checkpoint taxonomy:
FailureTaxonomy.md - Playwright-to-Python handoff contract:
Tools/CaptureStatusContract.md