/jaan-to:pm-research-about
Deep research on any topic, or add existing file/URL to research index.
What It Does
Two modes in one command:
- Deep Research — Give it a topic and it performs comprehensive 5-wave adaptive web research with parallel agents, then synthesizes findings into a structured markdown document.
- Add to Index — Give it a file path or URL and it extracts metadata, indexes it into the research collection, and commits.
The mode is auto-detected from your input.
Usage
/jaan-to:pm-research-about <topic-or-file-path-or-URL>
Examples:
/jaan-to:pm-research-about "Claude Code hooks best practices"— deep research/jaan-to:pm-research-about https://example.com/article— add URL to index/jaan-to:pm-research-about jaan-to/outputs/research/my-doc.md— add file to index
Mode Detection
| Input | Mode |
|---|---|
Starts with http:// or https:// | Add URL to index |
| Path exists as file | Add file to index |
| Anything else | Deep research |
Deep Research Mode
What It Asks
| Question | When |
|---|---|
| What topic? | If not specified |
| Clarifying questions (3-5) | If topic is broad or unclear |
| Research size? | Choose: 20/60/100/200/500 sources |
| Approval mode? | Choose: Auto/Summary/Interactive |
| Approve research plan? | Interactive mode only |
| Approve document? | Before saving (all modes) |
Research Sizes
| Size | Sources | Agents | Best For |
|---|---|---|---|
| Quick (20) | ~20 | 3 | Quick overview (3 waves) |
| Standard (60) | ~60 | 7 | Solid research (default) |
| Deep (100) | ~100 | 10 | Comprehensive |
| Extensive (200) | ~200 | 14 | In-depth analysis |
| Exhaustive (500) | ~500 | 29 | Maximum coverage |
Research Process
Uses 5 adaptive waves where each wave's focus is determined by findings from previous waves:
- Wave 1: Scout (1 agent) — Maps the research landscape
- Wave 2: Gaps (1-4 agents) — Fills primary gap from Scout
- Wave 3: Expand (1-10 agents) — Expands into new areas
- Wave 4: Verify (1-8 agents) — Verifies claims, resolves conflicts (size >= 60)
- Wave 5: Deep (2-6 agents) — Final deep dive (size >= 60)
Output
Path: jaan-to/outputs/research/{NN}-{category}-{slug}.md
Document sections: Executive Summary, Background, Key Findings, Recent Developments, Best Practices, Comparisons, Open Questions, Sources, Metadata.
Add to Index Mode
What It Asks
| Question | When |
|---|---|
| Approve proposal? | Before any changes |
| Category correct? | If auto-detected seems wrong |
For Local Files
- Reads first 50-100 lines
- Extracts title and summary
- Detects category
- Updates README.md index
For Web URLs
- Fetches content via WebFetch
- Creates
{NN}-{category}-{slug}.md - Updates README.md index
Categories
| Category | Keywords |
|---|---|
ai-workflow | claude, agent, workflow, prompt, MCP |
dev | code, architecture, testing, API |
pm | product, PRD, roadmap, feature |
qa | test, quality, CI/CD |
ux | design, UI, accessibility |
data | analytics, tracking, GTM |
growth | SEO, content, marketing |
mcp | MCP server, tool |
other | Fallback |
Also Does
- Updates
jaan-to/outputs/research/README.mdindex - Git commits the result
- Captures feedback via
/jaan-to:learn-add
Learning
This skill reads from:
jaan-to/learn/jaan-to:pm-research-about.learn.md
Add feedback:
/jaan-to:learn-add pm-research-about "your feedback here"