EcoTERRA

Ecological Tools for Evidence-based Research, Reproducibility & Analysis

A multi-agent method for quantitative ecology research — from literature discovery through data analysis, manuscript writing, and journal submission.

What is This?

EcoTERRA is a structured workflow that orchestrates specialized AI agents to assist with every phase of ecology research:

• Discovery — Literature search, biodiversity database exploration (GBIF, eBird, iNaturalist, IUCN)
• Strategy — Study design, power analysis, variable selection
• Data Engineering — Cleaning, taxonomic harmonization, environmental variable extraction
• Analysis — SDMs, occupancy models, community ecology, population dynamics, Bayesian modeling
• Writing — Manuscripts, supplementary materials, with journal-specific formatting
• Visualization — Publication-quality figures, maps, ordination plots
• Review — Simulated peer review (ecological domain + statistical methods + journal panel)
• Presentation — Conference talks, posters, science communication

Architecture

The core method lives in ecoterra/ and is model-agnostic — it works with any AI coding assistant. Tool-specific adapters (.claude/, .cursor/) wire the method into your IDE.

ecoterra/           ← The method (source of truth)
  agents/           ← 21 specialist agent personas
  workflows/        ← 13 slash-command workflows
  rules/            ← 8 domain rule sets
  knowledge/        ← Ecology domain knowledge base
  templates/        ← ODMAP, checklists, scaffolds

.claude/            ← Claude Code adapter (ships in git, works immediately)
.cursor/            ← Cursor adapter (generated via install.sh)

Quick Start

Option 1: npx (recommended)

# Start a new ecology research project
mkdir my-sdm-study && cd my-sdm-study
npx ecoterra init

# Or add EcoTERRA to an existing project
cd existing-project
npx ecoterra install

Option 2: Clone the repo

git clone https://github.com/drhammed/EcoTERRA.git
cd EcoTERRA
claude  # Start Claude Code — agents, commands, and rules are auto-detected

Option 3: Shell script (no npm required)

git clone https://github.com/drhammed/EcoTERRA.git
cd EcoTERRA
chmod +x install.sh
./install.sh                     # For Claude Code (default)
./install.sh --target cursor     # For Cursor users
./install.sh --target all        # Both

Workflows (Slash Commands)

| Command | Description | |---|---| | /new-project | Initialize a new ecology research project | | /discover | Literature search + biodiversity database exploration | | /strategize | Study design, power analysis, variable selection | | /wrangle | Data cleaning, harmonization, environmental variable extraction | | /analyze | Run statistical analysis pipeline | | /visualize | Generate publication-quality figures and maps | | /write | Draft or revise manuscript sections | | /review | Run adversarial peer review simulation | | /revise | Address reviewer comments | | /talk | Create conference presentation | | /submit | Pre-submission checklist | | /tools | Utility commands (compile, validate-bib, journal-format, reproducibility check) | | /checklist | Reproducibility and reporting standards check |

Agents

21 specialized agents organized in worker-critic pairs, plus infrastructure and review agents:

| Phase | Worker | Critic | |---|---|---| | Discovery | Librarian, Explorer | Librarian-Critic, Explorer-Critic | | Strategy | Strategist | Strategist-Critic | | Data | Data-Engineer | Data-Engineer-Critic | | Analysis | Analyst | Analyst-Critic | | Writing | Writer | Writer-Critic | | Visualization | Visualizer | Visualizer-Critic | | Presentation | Storyteller | Storyteller-Critic | | Review | EcoReviewer, MethodsReviewer, JournalReviewer | (adversarial by nature) | | Infrastructure | Orchestrator, Verifier | — |

The JournalReviewer simulates a full peer review panel (3 reviewers + editorial decision) calibrated to your target journal's standards.

Rules

8 domain rule sets, applied automatically based on file type:

| Rule | Scope | Purpose | |---|---|---| | contractor-mode | Always-on | Plan-first workflow, quality gates | | ecology-conventions | Always-on | Terminology, reporting standards | | r-code | *.R, *.Rmd | Tidyverse style, ecology packages | | python-code | *.py, *.ipynb | PEP 8, ecology Python stack | | data-rules | Data/** | FAIR principles, immutable raw data | | spatial-rules | *.shp, *.tif, *.gpkg | CRS, projections, resolution | | latex-rules | *.tex, *.bib | Journal templates, BibTeX | | quarto-rules | *.qmd | Manuscript/presentation conventions |

Quality Gates

| Score | Action | |---|---| | < 80 | Blocks commits | | 80–89 | Blocks pull requests | | 90–94 | Merge-ready | | 95+ | Excellence |

Who Is This For?

• PhD students — Structured workflow from messy data to submitted manuscript
• Early-career researchers — Quality gates catch common mistakes before submission
• Lab PIs — Reproducibility standards enforced automatically
• Conservation practitioners — Plain-language summaries from technical analyses
• Ecology instructors — Fork for course assignments with built-in review

Ecology-Specific Features

• Built-in knowledge of GBIF, eBird, iNaturalist, IUCN APIs and their data quality quirks
• Environmental data integration: WorldClim, CHELSA, ERA5, SoilGrids, MODIS, Google Earth Engine
• ODMAP protocol compliance for species distribution models
• Spatial awareness (CRS, scale, extent/resolution tradeoffs) across all agents
• Detection probability enforcement — flags analyses that ignore imperfect detection
• R-first design (ecology's lingua franca) with Python support
• Journal formatting for BES, ESA, Wiley, Nature, Elsevier ecology journals
• Simulated peer review calibrated by journal tier (Nature Eco Evo → PeerJ)
• FAIR data principles and Dryad/Zenodo/Figshare archiving workflows
• Reproducibility checklist, data management plan, and session persistence

Knowledge Base

EcoTERRA ships with ecology domain knowledge in ecoterra/knowledge/:

• Journals — Scope, word limits, formatting requirements for 20+ ecology journals
• Databases — GBIF, eBird, iNaturalist, IUCN, BioTIME, PREDICTS, WorldClim, CHELSA, ERA5, GEE
• R packages — 80+ ecology R packages organized by domain
• Python packages — Ecology Python stack
• Statistical methods — Decision guide for SDMs, occupancy, community, population, Bayesian
• Reporting standards — ODMAP, FAIR, TOP guidelines, statistical reporting conventions

Inspired By

• claude-code-my-workflow — Academic workflow template for Claude Code
• CLO-Author — AI research assistant for applied econometrics

License

MIT