scientific-brainstorming

Trigger phrases

Phrases that activate this skill when typed to Claude Code:

• brainstorm research ideas
• what are the open questions in
• explore research directions
• identify gaps in this field
• interdisciplinary connections for

What it does

scientific-brainstorming is a Claude Code skill from K-Dense AI’s scientific-agent-skills repo. It turns Claude into a research ideation partner for open-ended exploration — surfacing interdisciplinary connections, challenging underlying assumptions, and mapping the open questions in a field before you have specific experimental observations to work from.

A session produces a structured ideation output: a map of research directions, a list of under-explored questions, interdisciplinary connections that aren’t well-covered in the current literature, and explicit challenge questions for assumptions baked into the field’s consensus.

When to use it

Reach for it when:

• You’re entering a new research area and want to quickly map the conceptual landscape and identify where gaps exist
• You’re stuck in a local thinking pattern and want to force exposure to adjacent fields and non-obvious connections
• You’re planning a research program and want to pressure-test the premise before committing to specific experiments

When not to reach for it:

• You already have data and need to formulate testable hypotheses from it — use hypothesis-generation
• You need evidence-graded assessment of a specific claim — use scientific-critical-thinking

Install

Copy the SKILL.md from K-Dense AI’s scientific-brainstorming folder into .claude/skills/scientific-brainstorming/ in your project.

Trigger phrases: “brainstorm research ideas”, “what are the open questions in”, “explore research directions”, “identify gaps in this field”.

What a session looks like

A typical session has three phases:

1. Framing. You describe the research domain and any constraints (resources, expertise, timeline). Claude asks a small number of clarifying questions to avoid generating irrelevant directions.
2. Divergent exploration. Claude generates a broad map of research directions — conventional extensions, contrarian alternatives, and interdisciplinary imports from adjacent fields. Assumptions baked into the field’s standard framing are surfaced and challenged.
3. Structured output. The exploration is organized into a ranked set of research directions with brief rationale for each, plus a set of open questions organized by tractability and potential impact.

Receipts

Where it works well:

• Early-stage lab planning where the goal is “what should we work on next” rather than “how do we do X”
• Identifying interdisciplinary connections — Claude’s breadth across fields surfaces imports that a domain specialist might not encounter in their normal reading

Where it backfires:

• Without grounding in current literature, some generated “open questions” may already be answered in recent papers — follow up with literature-review to check
• The skill generates ideas, not feasibility assessments; resource and expertise constraints need human judgment applied after the session

Pattern that works: run a brainstorming session, pick the 2–3 most interesting directions, then immediately run literature-review on each to check what’s already been done — the combination takes under an hour.

Source and attribution

Originally authored by K-Dense Inc.. The canonical SKILL.md lives in the scientific-brainstorming folder of their public scientific-agent-skills repository.

License: MIT. Install, adapt, and redistribute with attribution preserved.

This page documents the skill from a practitioner’s perspective. For the formal spec and any updates, defer to the source repo.