By name: discovery description: Research market demand for marketing prompts/AI automation use cases. Find what's most valuable, in-demand, and fully automatable. Store findings in Researches/discovery/. user-invocable: false
Discovery Skill
Research what marketing prompts and AI automation use cases are most valuable right now
When to Run
Before deciding what to create or improve in the repo. Discovery informs priority — don't guess what's valuable, research it.
Research Process
1. Market Demand Research
Web search for current demand signals:
"most useful AI marketing prompts {current_year}""marketing tasks AI agents automate {current_year}""ChatGPT marketing use cases most popular""AI marketing automation workflows""marketing prompt engineering best practices {current_year}""what marketing tasks can be fully automated AI"
2. Automation Potential Research
Focus on what can run end-to-end without humans:
"AI agent marketing no human in the loop""fully automated marketing workflows AI""marketing tasks AI replaces completely""autonomous marketing agent use cases"
3. Competitor Analysis
Check what other prompt libraries prioritize:
"marketing prompt library github""AI marketing prompt collection best""marketing AI templates most starred github"
4. Trend Detection
What's emerging and underserved:
"AI marketing trends {current_year} new""marketing AI capabilities nobody uses""underrated AI marketing use cases"
What to Capture
For each finding, record:
| Field | Description |
|---|---|
| Use case | What the prompt/workflow does |
| Demand signal | How you know it's in-demand (search volume, mentions, stars, etc.) |
| Automation score | 1-5: can it run fully autonomous? (5 = no human needed) |
| Repo coverage | Does the repo already have this? Quality level? |
| Priority | High / Medium / Low based on demand x automation x gap |
Storage
Write findings to Researches/discovery/:
Researches/discovery/market-demand.md— Current demand signals, ranked use casesResearches/discovery/automation-candidates.md— Use cases scored by automation potentialResearches/discovery/competitor-analysis.md— What other libraries do well/missResearches/discovery/trends.md— Emerging opportunities
Each file should have a Last updated: YYYY-MM-DD header. Append new findings, don't overwrite — build a rolling knowledge base.
How to Use Findings
When deciding what to create/improve in the repo-improver workflow:
- Read
Researches/discovery/first - Cross-reference with repo gaps (empty/sparse directories)
- Pick the highest-impact action: high demand + high automation + low/no coverage = top priority