Yes. It is built for lawful cannabis cultivation operations, not generic gardening or broad-acre agriculture. Its workflows cover crop planning, controlled-environment incidents, root-zone and irrigation analysis, transparent calculations, harvest readiness, compliance verification, and operating records.
Read RELEASE-STATUS.md.
- Use the repository-native v0.1.5 line for the direct GitHub marketplace commands, root source tree, root tests, or repository-native Claude upload.
- Use the settled portable v0.1.6 bundle when you want one self-verifying archive with package-specific Codex and Claude payloads.
The v0.1.6 package preserves the same operating kernel; it does not create new field or behavioral evidence. Do not mix files across the two lines.
No. CanopyOps analyzes and documents information supplied through an AI host. It does not observe a facility by itself, perform physical work, operate equipment, apply products, submit compliance records, or release inventory.
No. It is an advisory reasoning-and-record system. Consequential recommendations remain proposals until an accountable human approves them, and approval remains distinct from execution and verification.
Not reliably enough to treat model memory as authority. CanopyOps helps identify what must be verified and how to record the result. Use current statutes, regulations, issuing regulators, facility policies, and qualified legal or compliance review.
It can organize observations and prepare product-permission questions. It must not invent cannabis applicability, jurisdiction permission, label directions, rates, intervals, worker-protection requirements, or facility authorization. The current approved label and accountable program owner govern.
It can build and rank a differential, identify discriminating observations, recommend safe reversible containment, and document the incident. It cannot create laboratory confirmation or turn an image, symptom list, or single sensor into certainty.
No for the core reasoning workflow. Python enables deterministic calculations, normalization, linting, packaging, freshness checks, schema-subset validation, and the portable v0.1.6 verifier. Without Python, small calculations may be shown transparently and marked manual or unverified.
Markdown, CSV, and JSON are first-class outputs. Templates cover facility and crop profiles, crop plans, incidents, cultivation decisions, harvest reviews, compliance verification, CAPA, risk registers, room runbooks, crop walks, drying logs, and shift handoffs.
CanopyOps itself includes no account, telemetry, analytics, hosted service, connector, MCP server, hook, or automatic network request. Your AI host, model, selected tools, repository host, and storage location govern data handling. See DATA-AND-PRIVACY.md.
No. The Pages site is static documentation with no JavaScript, tracking code, remote fonts, or equipment connection. A successful Pages deployment proves only that the documentation site was published.
No public-directory claim is made. A v0.1.5 skills-only draft was created and passed automated scanning on July 20, 2026. Owner attestations, review submission, approval, publication, and discoverability were not recorded. See PLUGIN-DIRECTORY-SUBMISSION-v0.1.5.md.
OpenAI’s current general Plugin Directory documentation is at https://help.openai.com/en/articles/20001256-plugins-in-codex. Local GitHub installation and public directory availability are separate states.
Yes. You may include and commercially redistribute the authentic unmodified CanopyOps with attribution as CanopyOps by Collaborative Dynamics. Authored Augment material uses CC-BY-ND-4.0; Python scripts, tests, and machine-readable schemas use MIT. See LICENSE.md and TRADEMARKS.md.
You may make private adaptations of the authored material, but CC-BY-ND-4.0 does not permit distributing those adaptations. MIT-licensed software and schemas may be modified and redistributed under their license; modified products should use a distinct identity unless separately authorized. Upstream contributions are welcome under CONTRIBUTING.md.
The Python scripts, tests, and schemas are open source under MIT. The complete authored Augment is free to use and commercially redistribute without derivatives under CC-BY-ND-4.0, so the whole product should not be described as open source.
No field pilot is claimed.
The v0.1.5 line has deterministic release checks and a reviewed three-case context-only safety/scope smoke inherited from the unchanged operating kernel. The v0.1.6 portable line adds static package and byte-custody evidence. Neither establishes field fitness, broad behavioral reliability, or customer outcomes.
Both distribution lines include packaged Claude skill surfaces. Their structures and adapters are documented, but current live upload, activation, invocation, progressive loading, script execution, and persistence have not been recorded as CanopyOps evidence.
For current Claude skill setup, see:
- https://support.claude.com/en/articles/12512180-use-skills-in-claude
- https://support.claude.com/en/articles/12512198-how-to-create-custom-skills
Report broken installation, missing files, incorrect calculations, schema or script failures, unsafe routing, invented authority, inaccessible documentation, stale source claims, version confusion, or behavior that contradicts the stated trust boundary. See SUPPORT.md.