solaris-api - SKILL.md Agent Skill

name: solaris_api description: Solaris Energy Forecasting API for off-grid community analysis and planning.

Solaris API Skill

Overview

Solaris is a multi-agent AI system that forecasts off-grid energy requirements for remote, unelectrified villages. The API is running at http://localhost:8000.

How to Use

Analyze a new location

Send a POST to create a location and trigger the full analysis pipeline:

curl -X POST http://localhost:8000/api/locations \
  -H "Content-Type: application/json" \
  -d '{"name": "Village Name", "lat": -1.286, "lon": 36.817, "households": 120}'

This returns {loc_id, name, run_id}. The pipeline runs perception (weather, demographics, seismic), spatial VLM (Sentinel-2 satellite imagery, NDVI, NDWI), and energy optimization agents in parallel.

Get location analysis results

curl http://localhost:8000/api/locations/{loc_id}

Returns the full pipeline output including:

outputs.demand_forecast — kWh/day with confidence intervals
outputs.scenario_set.primary — PV kW, battery kWh, solar kits
outputs.impact_metrics — CO2 avoided, cost savings, households served
outputs.quality — confidence score and status
runtime.agent_steps — step-by-step pipeline trace

Get satellite imagery data

curl http://localhost:8000/api/locations/{loc_id}/satellite

Returns NDVI, NDWI, SCL quality, vegetation change detection, and true-color Sentinel-2 preview URL.

Re-analyze a location

curl -X POST http://localhost:8000/api/locations/{loc_id}/reanalyze

List all monitored locations

curl http://localhost:8000/api/locations

Get dashboard aggregate stats

curl http://localhost:8000/api/dashboard/stats

Returns total locations, households, runs, and average confidence.

Direct pipeline execution (OpenClaw endpoint)

curl -X POST http://localhost:8000/openclaw/execute \
  -H "Content-Type: application/json" \
  -d '{"message": "Analyze energy needs", "lat": -1.286, "lon": 36.817, "households": 120}'

Returns full run payload plus loc_id and satellite, so frontend dashboards can open the stored result.

List OpenClaw execution history

curl "http://localhost:8000/api/openclaw/executions?limit=20"

Returns recent OpenClaw chat/execute logs with run_id, loc_id, status, and response_text.

Dashboard

The web dashboard is at http://localhost:5173 showing:

Interactive map with all monitored locations
Satellite imagery viewer (True Color, NDVI, NDWI)
Impact metrics, demand forecasts, energy blueprints
Agent pipeline trace and deployment timelines

Response Format

All agent outputs follow the shared guardrail contract: status, confidence, assumptions, quality_flags, provenance.