aquaponics-engineer

name	aquaponics-engineer
description	Compute cited, validated aquaponics and hydroponics system designs — tank/reservoir sizing, fish count, feed, pump, biofilter, bill of materials, operating envelope, and the best fish/crop ratio under a water budget. Deterministic engineering math, not an LLM guess: every number traces to a sourced coefficient and each result lists what it does NOT model. Use whenever a user asks to design, size, or optimize an aquaponic or hydroponic food system.
license	MIT
homepage	https://github.com/Rekin226/Agronaut
compatible	["claude-code","hermes-agent","openclaw"]

Aquaponics Engineer

A deterministic sizing engine for soil-less food systems, packaged as a portable skill. It wraps Agronaut's aqua_model trust zone — pure Python, no LLM, no network — so any agent can hand a user a computed, cited design instead of a plausible-sounding guess.

When to use

Invoke this skill when the user wants to design, size, or optimize an aquaponic (fish + plants) or hydroponic (plants only) system — e.g. "how big a tank for 12 m² of lettuce with tilapia?", "size a hydroponic bed for 10 m²", "what fish/crop ratio maximizes food from 5000 L/day?".

Do not invent sizing numbers yourself. Call the CLI and relay its output, including the cited coefficients and the "NOT modeled" caveats — that honesty is the point.

How to use

Run the CLI (Python 3.11+; pip install -r requirement.txt from the Agronaut repo):

# Aquaponics (fish + plants)
python -m skills.aquaponics_engineer.cli size-aquaponics \
    --fish tilapia --crop lettuce --area 12 --temp 27 --water 3000

# Hydroponics (plants only, nutrients dosed)
python -m skills.aquaponics_engineer.cli size-hydroponics \
    --crop lettuce --area 10 --temp 22 --water 500

# Best fish/crop ratio under a constraint
python -m skills.aquaponics_engineer.cli optimize \
    --area 10 --temp 28 --water 5000 --objective food     # food | protein | water_efficiency

# What's supported
python -m skills.aquaponics_engineer.cli list

Each command prints a compact, cited result. A bad input (unknown species/crop, out-of-range value) is rejected at the trust gate with VALIDATION_FAILED and a non-zero exit — ask the user for a corrected value rather than working around it.

What you get

• Sizing: tank/reservoir volume, fish count & biomass, feed/day, pump turnover, biofilter media (aquaponics), ET-driven water use and EC/nutrient target (hydroponics), makeup water.
• Feasibility against a daily water budget, with a nearest-feasible hint when it fails.
• A bill of materials and an operating envelope (pH, temperature, DO, EC bands).
• Every coefficient used, with its value, range, and source (FAO 589, UVI/Rakocy, literature) — and an explicit list of what the model does not cover.

Guarantees

• Deterministic: same inputs → same output. No randomness, no model drift.
• Auditable: every number traces to a cited coefficient.
• Honest: results state their own limits; seeds are calibration starting points, not universal constants.