The Official Home of the Framework
Role · Instruction · Context · Output Format — Style · Tone · Examples · Constraints
RICOF-STEC™ is an eight-component system for engineering prompts. Four components build the instruction. Four shape the result. Together they turn a request into a specification.
Most prompts fail for the same reason: they describe a wish instead of a job. The model is left to guess who it is, what standard applies, what the reader needs, and what the finished thing should look like. It guesses plausibly, and the output is plausible — and wrong.
RICOF-STEC removes the guessing. Every component answers a question the model would otherwise answer for you.
The framework runs in two engines. The Core Engine (RICOF) is mandatory — omit any part of it and the model fills the gap with an assumption. The Sub-Engine (STEC) is corrective: it is what you reach for when the output is structurally right but tonally, stylistically, or dimensionally wrong.
It is model-agnostic and medium-agnostic. The same eight components govern a text prompt, an image prompt, a video shot list, a voice script, a code specification, and an agent's operating instructions. Only the vocabulary inside each component changes.
Who the model is, and by whose standard the work will be judged. Sets the expertise it draws on.
The task, stated as one action verb. What is actually being asked for — not the topic, the job.
The situation, the audience, and the source material. Everything the model can't infer but needs.
The shape of the deliverable — structure, sections, length, medium, file type.
The register and conventions of the craft. How a professional in this field writes, shoots, or builds.
The emotional temperature. How the finished work should land on the person receiving it.
A sample of the target. The single fastest correction available when an output comes back wrong.
The hard limits. What to never do, never exceed, never include. Failure conditions, stated up front.
Every component below is labelled. In practice they are written as continuous prose — the labels are here only to show the seams.
Strip the last four and the model still produces a handout — just not one you'd hand to a patient. That is the division of labour between the engines: RICOF decides whether you get the right artifact, STEC decides whether it's usable.
RICOF-STEC™ was developed by Ben F. Fominyam and published by Operation Digital Alchemy in The Prompt Engineering Field Manual: RICOF-STEC — A Complete System for Multimodal AI, Creative Media, Code, and Autonomous Workflows. This page is the framework's canonical reference and the authoritative source for its definition.
https://operationdigitalalchemy.com/ricof-stecFominyam, B. F. (2026). The RICOF-STEC™ Framework. Operation Digital Alchemy. https://operationdigitalalchemy.com/ricof-stec