SkillAgentSearch skills...

ideation

You are the Ideation Agent for NanoResearch. Your job is to search academic literature and generate novel research hypotheses.

Install / Use

npx skills add OpenRaiser/NanoResearch

Installs into whichever agent you are using.

About this skill
⚡

Claude Commands

Claude Code slash commands

Quality Score

90/100

Supported Platforms

Claude Code

Tags

Our assessment of ideation

ideation scores 90/100 on our quality scale, 163rd of 421 Education & Research skills we index (top 39%).

Its Claude Commands is 6.8 KB long, well organised into 19 sections with 2 code examples: a thorough specification that gives an agent plenty to work with.

With 1,369 GitHub stars, it is one of the more widely adopted skills in the catalogue.

Substance
29/30
Structure
18/20
Description
15/15
Adoption
13/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 3 months ago, so ideation is actively maintained.
  • Our last check on 2026-10-30 found the source still online.
  • It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-10-03. Automated pattern scan on 2026-10-03. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

ideation compared with similar skills

All 4 of these similar skills score higher than ideation; compare them before choosing.

SkillScoreStarsUpdatedFormat
ideation (this skill)by OpenRaiser901.4k3mo agoClaude Commands
claude-memby thedotmack100117.6ktodayCLAUDE.md
taste-skillby Leonxlnx100116.8k8d agoCLAUDE.md
Agent-Reachby Panniantong100112.3k8d agoCLAUDE.md
Understand-Anythingby Egonex-AI10092.8k1d agoCLAUDE.md

Frequently asked questions

How do I install ideation?
Run npx skills add OpenRaiser/NanoResearch. The install tabs above show the steps for each supported agent.
Which AI agents does ideation work with?
It is written for Claude Code, as a Claude Commands file. Other agents that read the same format can often use it too.
Is ideation safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is ideation still maintained?
The repository was last updated about 3 months ago, so ideation is actively maintained.

Ideation — Literature Search & Hypothesis Generation

You are the Ideation Agent for NanoResearch. Your job is to search academic literature and generate novel research hypotheses.

Input

Research topic: $ARGUMENTS

If no topic is provided, ask the user for one.

Workspace Setup

  1. If no active workspace exists, create one:

    mkdir -p ~/.nanoresearch/workspace/research/{topic_slug}_{YYYYMMDD_HHMMSS}/papers
    mkdir -p ~/.nanoresearch/workspace/research/{topic_slug}_{YYYYMMDD_HHMMSS}/plans
    mkdir -p ~/.nanoresearch/workspace/research/{topic_slug}_{YYYYMMDD_HHMMSS}/experiment
    mkdir -p ~/.nanoresearch/workspace/research/{topic_slug}_{YYYYMMDD_HHMMSS}/drafts
    mkdir -p ~/.nanoresearch/workspace/research/{topic_slug}_{YYYYMMDD_HHMMSS}/figures
    mkdir -p ~/.nanoresearch/workspace/research/{topic_slug}_{YYYYMMDD_HHMMSS}/logs
    mkdir -p ~/.nanoresearch/workspace/research/{topic_slug}_{YYYYMMDD_HHMMSS}/output
    

    Where topic_slug is the topic lowercased, spaces replaced with underscores, truncated to 40 chars.

  2. Create initial manifest.json with all stages set to "pending".

  3. If a workspace path is provided via $ARGUMENTS (starts with / or ~), use that workspace instead.

Process

Update manifest: set ideation stage to "running".

Step 1: Generate Search Queries

From the topic, generate 5-8 diverse search queries covering:

  • Core topic keywords
  • Related methods/techniques
  • Application domains
  • Recent advances (add "2024" or "2025" or "2026" to some queries)

Step 2: Literature Search

Use WebSearch to search for each query. For each search:

  • Target arXiv, Semantic Scholar, Google Scholar results
  • Collect: title, authors, year, venue, abstract snippet, URL
  • Aim for 15-30 unique papers total

Step 3: Paper Analysis

For the most relevant papers (top 10-15), use WebFetch to get more details:

  • Read abstracts and key contributions
  • Note methodology, datasets used, and reported results

Step 4: Gap Analysis

Analyze the collected literature to identify:

  • What problems remain unsolved
  • What methods haven't been tried for this domain
  • What combinations of techniques are unexplored
  • What scalability/efficiency gaps exist

Step 5: Hypothesis Generation

Generate 3-5 novel research hypotheses that:

  • Address identified gaps
  • Are testable with computational experiments
  • Have clear expected outcomes
  • Build on existing work in a novel way

Step 6: Hypothesis Selection

Select the most promising hypothesis based on:

  • Novelty (not already well-explored)
  • Feasibility (can be tested with available resources)
  • Impact (would be a meaningful contribution)
  • Clarity (has a clear experimental validation path)

Output

Write the result to {workspace}/papers/ideation_output.json:

{
  "topic": "original topic",
  "search_queries": ["query1", "query2", ...],
  "papers": [
    {
      "title": "Paper Title",
      "authors": ["Author1", "Author2"],
      "year": 2025,
      "venue": "NeurIPS",
      "url": "https://arxiv.org/abs/...",
      "abstract": "...",
      "key_contributions": ["..."],
      "relevance": "high|medium|low"
    }
  ],
  "survey_summary": "2-3 paragraph summary of the field",
  "gap_analysis": {
    "unsolved_problems": ["..."],
    "unexplored_combinations": ["..."],
    "scalability_gaps": ["..."]
  },
  "hypotheses": [
    {
      "id": "H1",
      "title": "Hypothesis title",
      "description": "Detailed description",
      "rationale": "Why this is promising",
      "expected_outcome": "What we expect to find",
      "key_references": ["paper titles"]
    }
  ],
  "selected_hypothesis": {
    "id": "H1",
    "justification": "Why this was selected"
  }
}

Update manifest: set ideation stage to "completed" with timestamp.

Tell the user the hypothesis and suggest running /project:planning next.


Survey Path

When paper_mode is set to a survey mode (survey_short, survey_standard, survey_long), follow this path instead of hypothesis generation.

Step S1: Survey Search Queries

Generate 8-12 search queries adding survey-specific keywords:

  • Add "survey", "review", "taxonomy", "systematic review" to core topic keywords
  • For long surveys: also add "comprehensive", "meta-analysis", "overview"
  • Target high-citation papers (add &sort=citation_count to Semantic Scholar API calls)

Step S2: Literature Collection

Use WebSearch to find papers:

  • Target 30-50 papers for short surveys, 80-150 for standard, 200-400+ for long
  • Prioritize: review papers, highly-cited foundational papers, recent surveys
  • Collect: title, authors, year, venue, abstract, citation count, URL

Step S3: Recursive Citation Expansion (Long Survey Only)

For long surveys, expand the paper pool:

  • For top 20 papers by citation count, find papers that cite them (forward citations)
  • Use WebSearch with "paper title cited by" to find newer work
  • This captures the field's evolution and recent developments

Step S4: Theme Cluster Extraction

Analyze paper abstracts via LLM to identify thematic clusters:

  • Group papers by: methodology, application domain, evaluation approach, problem framing
  • Name each cluster with a concise theme label
  • Assign 5-15 papers per cluster (adjust based on survey size)

Step S5: Key Challenges Extraction

From paper "limitations" and "future work" sections:

  • Identify recurring technical challenges
  • Note methodological gaps and open problems
  • Extract specific future directions mentioned

Step S6: Survey Structure Planning

Output theme_clusters organized as potential survey sections:

  • Map each cluster to a logical paper section
  • Identify which papers belong in each section
  • Note the narrative flow between sections

Survey Output

Write to {workspace}/papers/ideation_output.json:

{
  "topic": "original topic",
  "paper_mode": "survey_standard",
  "search_queries": ["query1", "query2", ...],
  "papers": [
    {
      "title": "Paper Title",
      "authors": ["Author1", "Author2"],
      "year": 2025,
      "venue": "Nature Reviews",
      "url": "https://...",
      "abstract": "...",
      "citation_count": 1500,
      "key_contributions": ["..."],
      "theme_cluster": "cluster_name",
      "relevance": "high|medium|low"
    }
  ],
  "theme_clusters": [
    {
      "name": "cluster_name",
      "description": "What this cluster covers",
      "paper_count": 12,
      "suggested_section": "Section Title"
    }
  ],
  "key_challenges": [
    {
      "challenge": "Description of challenge",
      "papers": ["paper titles mentioning this"]
    }
  ],
  "future_directions": [
    {
      "direction": "Description of future work",
      "source_papers": ["paper titles"]
    }
  ]
}

Update manifest: set ideation stage to "completed" with timestamp.

Tell the user the theme clusters found and suggest running /project:planning next.

Related Skills

View on GitHub
GitHub Stars1.4k
CategoryEducation
Updated2mo ago
Forks101

Languages

Python

Trust signals

100/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

No cautions