| name | suggest-research |
| description | Review repository evidence and propose high-impact next research directions for EcoSIM, emphasizing warming-nitrogen mechanisms, data readiness, and methodological feasibility. |
| allowed-tools | Bash, Read, Write, Edit, WebSearch, AskUserQuestion |
| user-invocable | true |
Suggest Research Skill
Identify and prioritize the next research topic using available EcoSIM-relevant data and completed analyses in this repository.
Compatibility Modes
- Full mode: BERDL data access plus MCP literature tooling available.
- Limited mode: rely on repository-local evidence and web literature fallback.
Usage
/suggest-research
Workflow
Step 1: Inventory Available Evidence
Read and summarize key inputs:
derived/*.tsv datasets
derived/*/summary.md and derived/*/metadata.yaml
bervo/bervo-terms.tsv
hackathon-case_study-experimental_warming_nitrogen/derived/*.tsv
PLANT_TRAITS.md and hackathon-case_study-experimental_warming_nitrogen/plant_traits.json
Step 2: Extract Reusable Findings and Gaps
From summaries and metadata, capture:
- Repeated findings across ecosystems
- Common missing variables
- Methodological limitations
- Sites or ecosystems with sparse coverage
Step 3: Assess Data Readiness
For each candidate topic, check:
- Do we have benchmark observations?
- Do we have required climate/soil/trait inputs?
- Is BERVO mapping available for required variables?
- Can EcoSIM plausibly simulate the process?
Step 4: Optional User Constraints
If the user has not specified preferences, ask:
- Theme preference (for example: nitrification, denitrification, soil moisture)
- Effort preference (low/medium/high)
- Extension of existing analyses vs novel direction
Step 5: Score Candidate Topics
Use this rubric:
| Criterion | Weight |
|---|
| Scientific novelty | High |
| Data readiness | High |
| Potential impact | High |
| Feasibility | Medium |
| Reuse of existing assets | Medium |
| Effort fit | Low |
Pick one primary recommendation plus one alternative.
Step 6: Novelty Check
Run literature-review for the top candidate to confirm:
- What is already known
- What is unresolved
- Where this repository can contribute uniquely
Step 7: Present Recommendation
Provide:
- Topic title
- Research question
- Hypotheses (H1, H0, optional H2)
- Why now
- Required datasets and variables
- Planned approach
- Expected impact
- Risks and dependencies
- Effort estimate
- Alternative topic
Step 8: Optional Backlog Update
If user approves, append to docs/research_ideas.md (create if missing) with status PROPOSED.
Do not require BERIL-specific project scaffolding commands.
Integration
- Reads from: repository data products and summaries
- Calls:
literature-review for novelty checks
- Optionally writes:
docs/research_ideas.md
Pitfall Handling
If blocked by missing inputs or contradictory metadata, record concise notes in repo memory and surface assumptions in the recommendation.