This skill should be used when the user asks to "research a topic", "create a research doc", "explore options for X", "what are the different approaches to X", "write up research on X", or needs a broad landscape scan on any subject. Produces a concise, well-segmented .md research file covering a diverse range of options, approaches, or findings.
Create a well-segmented markdown research file that covers a topic broadly and diversely. Writing is parsimonious — no filler, no padding, every sentence earns its place.
Ask the user:
Before writing, brainstorm the broadest possible landscape of segments. The goal is diversity and coverage — not depth yet.
Segmentation strategies to consider (use whichever fit the topic):
Present the proposed segments to the user for confirmation before writing. Aim for 6-12 clean, non-overlapping segments that collectively cover the full landscape.
After segments are confirmed, identify 6–10 (or more) criteria that matter for evaluating and comparing these segments. These criteria will be used by the analyzer skill to score and rank segments.
Criteria should be:
Present the proposed criteria to the user for confirmation. Include them in the research file under a dedicated section so the analyzer has them ready.
Structure:
# [Topic] Research
**Author:** [Name]
**Date:** [YYYY-MM-DD]
## Overview
[2-3 sentences framing the topic and why it matters. State the research question.]
## Landscape
[One paragraph summarizing the full range of what exists. Set up the segments.]
## [Segment 1 Name]
**What it is:** [1-2 sentences]
**Key players/examples:** [Names, links if available]
**Strengths:** [Bullets]
**Weaknesses:** [Bullets]
**Relevance to Pocket Space:** [1 sentence — how this connects]
## [Segment 2 Name]
[Same structure]
...
## Proposed Evaluation Criteria
[6-10 criteria for comparing these segments. Each with a one-line rationale.]
| # | Criterion | Why It Matters |
|---|---|---|
| 1 | [Name] | [One-line rationale] |
| 2 | [Name] | [One-line rationale] |
| ... | ... | ... |
## Key Takeaways
- [3-5 bullets distilling the most important findings]
## Open Questions
- [What this research didn't answer]
- [What needs deeper investigation]
## Sources
- [List of sources used]
Research/ subfolder already exists for this topic, place the .md file inside Research/ (or a nested subfolder within it)[Department_Folder]/
└── [Topic_of_Research]/
└── Research/
├── README.md
└── [Topic]_Research.md ← This file goes here
The .docx will be added later by the research-synthesizer skill at the top level of the topic folder.[Topic]_Research.md using underscore conventionAll research files use Vancouver citation style — numbered references in order of first appearance.
Inline: Use superscript or bracketed numbers in the text where a claim is made.
Bonding curves were first proposed in 2017 [1] and later adopted in DeFi protocols [2].Reference list: A numbered ## References section at the end of the document.
[Number] Author(s). Title. Source/Publication. Year. URL (if applicable).## References
[1] de la Rouviere S. Tokens 2.0: Curved token bonding. Medium. 2017. https://example.com
[2] Zargham M. An introduction to bonding curves. BlockScience. 2019. https://example.com
Rules:
[Author's analysis]Before finalizing segments, verify: