# Managing delegated AI work — OPL field guide v3

> Proposed framework for practitioner testing. Not practitioner-validated; fictional cases. Completeness checks do not verify evidence or authorise deployment.

## Read and inspect

- [Complete reference](https://opl-agentic-anatomy.netlify.app/reference.html): All explanatory material, fictional cases, evidence prompts, units and limitations; no JavaScript needed.
- [Interactive guide](https://opl-agentic-anatomy.netlify.app/): Local record builder and workload calculator.
- [Public reference JSON](https://opl-agentic-anatomy.netlify.app/guide.json): The same content used by the interactive guide.
- [Record JSON Schema](https://opl-agentic-anatomy.netlify.app/record.schema.json): Structure validation only; semantic checks remain necessary.
- [Blank record](https://opl-agentic-anatomy.netlify.app/blank-record.json): Local interchange template with unknown evidence and illustrative workload defaults.

## Context

- [More Open Work](https://davidseymour.uk/for-agents.html): related public projects and existing feedback routes.

No submission API or automatic record transmission. Cite OPL Advisory, field guide v3, section URL and access date. Sources and reuse terms retain their own attribution. The companion paper is not published here.