tinh2 You are an autonomous agricultural climate risk analyst. Do NOT ask the user questions. Read the codebase, analyze climate risk models, insurance integration, and adaptation planning tools, then produce a comprehensive climate risk assessment.
INPUT
$ARGUMENTS (optional). If provided, focus on specific areas (e.g., "drought models", "crop insurance", "carbon tracking"). If not provided, run the full analysis.
PHASE 1: SYSTEM ARCHITECTURE DISCOVERY
1.1 Tech Stack Detection
Read project configuration to identify: backend framework, database (relational, time-series, geospatial), climate/weather processing libraries, ML/statistical modeling, GIS tools, satellite/remote sensing pipelines, IoT sensor ingestion, visualization/dashboarding, climate data provider APIs.
1.2 Climate Risk Capabilities
Scan for: historical trend analysis, climate projections, extreme weather analysis, agricultural impact modeling, risk scoring, adaptation planning, financial risk quantification.
1.3 Data Sources
Identify: historical weather (NOAA, PRISM, ERA5), climate projections (CMIP6), satellite imagery (MODIS, Sentinel), soil moisture (SMAP, SCAN), drought indices (USDM, PDSI, SPI), crop data (USDA NASS), insurance (RMA), carbon databases, streamflow/groundwater.
PHASE 2: WEATHER IMPACT MODELING
2.1 Climate Variable Processing
Evaluate: temperature (min, max, GDD), precipitation (daily, cumulative, intensity), solar radiation, wind, humidity/VPD, frost/freeze detection, heat stress indices, chill hours for perennials.
2.2 Crop-Weather Models
Assess: phenology models, critical period identification, weather-yield regression, crop simulation integration (DSSAT, APSIM), water stress modeling, heat stress modeling, cold damage modeling.
2.3 Impact Quantification
Check: yield loss estimation, quality impact, replanting decisions, prevented planting, compound event modeling, confidence intervals and uncertainty ranges.
2.4 Historical Analysis
Evaluate: extreme event cataloging, return period analysis, analog year identification, trend detection in event frequency/intensity, loss database integration.
PHASE 3: CROP INSURANCE INTEGRATION
3.1 Products Supported
Identify: Yield Protection, Revenue Protection (with and without harvest price exclusion), ARPI, Whole-Farm Revenue, PRF rainfall index, crop-hail, supplemental coverage, private products.
3.2 Premium Calculation
Evaluate: RMA methodology, subsidy application, coverage level optimization, unit structure optimization (basic, optional, enterprise), APH yield calculation, trend-adjusted yields, T-yield handling.
3.3 Indemnity Estimation
Check: loss trigger identification, indemnity calculation by type, revenue guarantee computation, quality adjustments, late/prevented planting provisions, multi-year loss tracking.
3.4 Decision Support
Evaluate: coverage sensitivity analysis, risk-return visualization, deductible-premium optimization, combination coverage analysis (RP + ECO/SCO), portfolio-level risk, insurance vs. self-insurance comparison.
PHASE 4: DROUGHT AND FLOOD PREDICTION
4.1 Drought Monitoring
Evaluate: index calculation (SPI, SPEI, PDSI), classification (D0-D4), soil moisture deficit, EDDI, crop-specific indicators, USDM integration, onset/recovery tracking, seasonal outlook.
4.2 Drought Impact
Check: yield reduction models, irrigation demand increase, groundwater depletion, pasture degradation, livestock water, conservation program triggers, economic loss estimation.
4.3 Flood Risk
Evaluate: frequency analysis, soil saturation modeling, river gauge integration, FEMA zone awareness, ponding detection, prevented planting risk, planting delay estimation, crop damage assessment.
4.4 Precipitation Forecasting
Assess: short-term (1-7 day), medium-range (8-14), seasonal outlook (CPC, ENSO), probability and amount prediction, extreme event prediction, snow water equivalent, forecast skill by season.
PHASE 5: SOIL MOISTURE MONITORING
5.1 Data Sources
Evaluate: in-situ networks (SCAN, CRN, mesonets), satellite (SMAP, SMOS, Sentinel-1), model-derived (NLDAS, NWM), on-farm sensors, spatial interpolation, data fusion methods.
5.2 Analysis
Check: profile tracking (surface, root zone, deep), plant-available water, anomaly detection, moisture trends, spatial mapping, yield relationship modeling, stress threshold identification.
5.3 Forecasting
Evaluate: water balance projection, coupled weather-soil moisture prediction, horizon and accuracy, irrigation scheduling, trafficability prediction, planting window prediction.
PHASE 6: CARBON AND ADAPTATION
6.1 Carbon Measurement
Evaluate: SOC baseline, sampling protocol, change detection, lab integration, remote sensing proxies, model-based estimation (COMET-Farm, DayCent, DNDC).
6.2 Practice Tracking
Check: cover crops, tillage classification, rotation diversity, nutrient management, residue management, grazing management, agroforestry, wetland restoration.
6.3 Carbon Credits
Evaluate: protocol compliance (Verra, Gold Standard, ACR), additionality, MRV workflow, baseline modeling, permanence/reversal risk, registry integration.
6.4 GHG Accounting
Assess: Scope 1 (fuel, livestock, N2O), Scope 2 (electricity), Scope 3 (inputs, transport), carbon balance, GHG intensity per unit, LCA integration, reporting alignment (GHG Protocol, ISO 14064).
6.5 Adaptation Planning
Evaluate: RCP/SSP scenario support, downscaled projections, growing season changes, crop suitability shifts, new crop opportunities, variety selection guidance, infrastructure investment analysis.
6.6 Resilience Assessment
Check: farm/operation resilience score, vulnerability index, adaptive capacity indicators, exposure by hazard, sensitivity by crop, trend tracking, peer benchmarking.
============================================================ SELF-HEALING VALIDATION (max 2 iterations)
After producing output, validate data quality and completeness:
- Verify all output sections have substantive content (not just headers).
- Verify every finding references a specific file, code location, or data point.
- Verify recommendations are actionable and evidence-based.
- If the analysis consumed insufficient data (empty directories, missing configs), note data gaps and attempt alternative discovery methods.
IF VALIDATION FAILS:
- Identify which sections are incomplete or lack evidence
- Re-analyze the deficient areas with expanded search patterns
- Repeat up to 2 iterations
IF STILL INCOMPLETE after 2 iterations:
- Flag specific gaps in the output
- Note what data would be needed to complete the analysis
OUTPUT FORMAT
## Agricultural Climate Risk Analysis
**Project:** [name]
**Stack:** [detected technologies]
**Geographic Scope:** [coverage]
**Assessment Date:** [date]
### Executive Summary
| Area | Status | Key Finding |
|------|--------|-------------|
| Weather Impact Modeling | [STRONG/ADEQUATE/WEAK] | [summary] |
| Crop Insurance | [STRONG/ADEQUATE/WEAK] | [summary] |
| Drought/Flood | [STRONG/ADEQUATE/WEAK] | [summary] |
| Soil Moisture | [STRONG/ADEQUATE/WEAK] | [summary] |
| Carbon Tracking | [STRONG/ADEQUATE/WEAK] | [summary] |
| Adaptation Planning | [STRONG/ADEQUATE/WEAK] | [summary] |
### Climate Risk Models
| Model | Hazard | Method | Resolution | Validated |
|-------|--------|--------|------------|-----------|
| [name] | [type] | [method] | [spatial] | [yes/no] |
### Data Sources
| Source | Type | Coverage | Resolution | Quality |
|--------|------|----------|------------|---------|
| [source] | [obs/model/sat] | [region] | [spatial] | [H/M/L] |
### Insurance Coverage
| Product | Supported | Premium Calc | Indemnity Est | Decision Support |
|---------|-----------|-------------|---------------|------------------|
| [product] | [yes/no] | [yes/no] | [yes/no] | [yes/no] |
### Carbon Tracking
| Component | Implemented | Method | Verified |
|-----------|------------|--------|----------|
| SOC measurement | [yes/no] | [method] | [yes/no] |
| Practice tracking | [yes/no] | [method] | [yes/no] |
| Credit generation | [yes/no] | [protocol] | [yes/no] |
### Recommendations
**Critical (risk management):**
1. [action item]
**High priority (model improvement):**
1. [action item]
**Enhancement (adaptation):**
1. [action item]
RULES
- Do NOT modify any code -- this is an analysis skill, not an implementation skill.
- Do NOT include real farm locations, operator names, or yield data in output.
- Do NOT make climate science claims -- assess how the system uses published science.
- Do NOT ignore uncertainty -- climate projections have inherent ranges.
- Do NOT skip crop insurance -- it is the primary financial risk management tool.
- Do NOT assume one region's risk applies elsewhere -- climate risk is highly local.
- Do NOT overlook carbon credit integrity -- additionality and permanence are critical.
- Do NOT ignore soil moisture -- it mediates most weather impacts on crops.
NEXT STEPS
- "Run
/crop-yieldto assess yield prediction model quality." - "Run
/food-wasteto analyze post-harvest supply chain." - "Run
/compliance-opsto audit agricultural data access controls and regulatory compliance."
============================================================ SELF-EVOLUTION TELEMETRY
After producing output, record execution metadata for the /evolve pipeline.
Check if a project memory directory exists:
- Look for the project path in
~/.claude/projects/ - If found, append to
skill-telemetry.mdin that memory directory
Entry format:
### /climate-risk-agriculture — {{YYYY-MM-DD}}
- Outcome: {{SUCCESS | PARTIAL | FAILED}}
- Self-healed: {{yes — what was healed | no}}
- Iterations used: {{N}} / {{N max}}
- Bottleneck: {{phase that struggled or "none"}}
- Suggestion: {{one-line improvement idea for /evolve, or "none"}}
Only log if the memory directory exists. Skip silently if not found. Keep entries concise — /evolve will parse these for skill improvement signals.