SYSMOD

Why a story

Why tell this as a story at all?

Each step that follows is told five times: as a story, as a formal definition, as a SysML v2 library element, as an example model, and as an AI skill. The story comes first because a method is easier to remember, teach, and defend in front of stakeholders when it’s carried by a concrete narrative rather than an abstract process description. Here, that narrative follows a fictional delivery-drone project at a company called Amperdrones — simple on purpose, so the sequence of thinking stays memorable rather than realistic.

What SYSMOD is

A brief definition, and where it sits in ISO 15288

SYSMOD is a pragmatic, lightweight method for Model-Based Systems Engineering: a focused chain of artifacts, from problem statement to verification, that keeps a project traceable without demanding a full heavyweight process.

The method itself is language-independent. It defines what each artifact is and how the artifacts relate, not how they are written down in a particular notation. A separate language mapping binds each SYSMOD concept to the constructs of a concrete modeling language. SYSMOD has also been applied with other modeling languages and in document-based environments without a formal model. This site documents only the mapping to SysML v2, the primary language of the current SYSMOD version.

ISO/IEC/IEEE 15288: SYSMOD is not a complete implementation of ISO/IEC/IEEE 15288. It touches several technical processes (Business or Mission Analysis, Stakeholder Needs & Requirements Definition, System Requirements Definition, Architecture Definition, Verification, Validation) but deliberately omits others, and does not implement every activity within the processes it does address.

The SysML v2 library

Where the real, versioned library lives

SYSMOD concepts are covered by a SysML v2 library. Each step page shows the library elements for that step. The full, versioned SYSMOD SysML v2 library — every definition, package, and import — is maintained openly on GitHub.

mbse4u/sysmod-sysmlv2 (GitHub)

About the example

Oversimplified on purpose

The delivery drone example is deliberately oversimplified: a handful of stakeholders, some requirements, some use cases, and so forth. Real projects have much more. The simplification keeps the mapping between story, definition, and model easy to follow rather than realistic.

The AI skills

Delegating a step to an AI assistant

Every main artifact in the SYSMOD SysML v2 library carries AI metadata: a machine-readable skill with a name, create_questions, a create_prompt, and a validation_prompt, plus a perform_prompt that runs all three as one ready-to-execute workflow. AIProject attaches one of these to each main artifact of Project, in SYSMOD order; an assistant matches a given artifact’s AI metadata (e.g. problemStatementAI) onto the identically-named feature of your own project — a real project stays a plain Project and never specializes AIProject. Two further prompts, story_prompt and slide_prompt, let an assistant explain any selected element to non-technical stakeholders. Each skill’s own doc comment also spells out exactly when to use it and when not to — the same trigger/skip framing an AI-assistant skill would use — so these elements can be turned directly into Claude Skills or equivalent definitions for other LLM platforms; see the full library for the complete text. Each step page shows that step’s skill, condensed from the library.

metadata def AI {
    attribute name : ScalarValues::String {
        doc
        /* Unique identifier of the AI skill, for example, 'sysmod-problem-statement'. */
    }
    attribute create_prompt : ScalarValues::String {
        doc
        /* Step-by-step instructions for an AI assistant to create the annotated artifact. */
    }
    attribute create_questions : ScalarValues::String [*] {
        doc
        /* Questions the AI assistant can ask the user to elicit missing information. */
    }
    attribute validation_prompt : ScalarValues::String {
        doc
        /* Instructions for an AI assistant to verify and validate the annotated artifact against the SYSMOD rules. */
    }
    // plus perform_prompt, which runs create and validate as one workflow,
    // and story_prompt, slide_prompt for stakeholder-facing output
}
// e.g., inside AIProject, matched onto your project by feature identity:
metadata problemStatementAI : AI about problemStatement {
    // name, create_questions, create_prompt, validation_prompt, ...
}