What is canopyops?

CanopyOps is an AI skill built for cannabis cultivation operations. It is designed to work inside AI assistants like Codex and Claude, taking in observations, logs, and facility details from a grow operation and turning them into structured records: crop plans, diagnostic writeups, harvest reviews, incident reports, and shift handoffs. The goal is to make the reasoning behind a cultivation decision visible, including what evidence was used, what assumptions were made, and who still needs to approve something, rather than just spitting out an answer. With it, a grower could build a room and crop plan from facility limits and cultivar information, work through an environmental or pest problem step by step instead of jumping to one guess, calculate figures like vapor pressure deficit or irrigation volumes, review whether a harvest is actually ready, and generate paperwork such as CAPA reports and risk registers. Everything it produces is plain Markdown, CSV, or JSON files using included templates, so the output is meant to be readable and reusable rather than proprietary. It was built by a person working with Codex and GPT-5.6 during an OpenAI event in July 2026, with the human providing the product direction, source material, and judgment about what the tool should and should not do. The project explicitly states it is advisory only: it will not authorize pesticide use, interpret local law, approve inventory releases, or replace the facility's own procedures and accountable staff. It includes an 18-test suite so people can check how it behaves rather than take the claims on faith. CanopyOps supports installation as a Codex plugin, a standalone Codex skill, a Claude.ai custom skill upload, or a Claude Code skill, though the README notes that live use in some of these hosts has not yet been recorded. Python is optional and only needed for extra deterministic calculations and validation, the core workflow can run without it. The Augment content itself is licensed CC-BY-ND-4.0 for commercial reuse with attribution, while the Python scripts, tests, and schemas are under MIT.