Skip to content

spectralbranding/orgschema-toolkit

BranchesTags

Repository files navigation

MIT License CC-BY 4.0 Last Updated

Organizational Schema Theory — Prompt Toolkit

Businesses are test suites. You are the architect.

Toolkit v1.0 5 Industry Starters

graph TD
    subgraph Cascade["TDD Cascade (top-down design)"]
        L0["Module 1<br/>Experience Design<br/>(L0: Acceptance Tests)"]
        L1["Module 2<br/>Signal Architecture<br/>(L1: Integration Tests)"]
        L2["Module 3<br/>Process Contracts<br/>(L2: Unit Tests)"]
        L3["Module 4<br/>Procedures<br/>(L3: Implementation)"]
        L4["Module 5<br/>Inputs<br/>(L4: Dependencies)"]
        L5["Module 6<br/>Sourcing<br/>(L5: Infrastructure)"]
    end

    subgraph Validation["Validation"]
        M7["Module 7<br/>Signal Mapping"]
        M8["Module 8<br/>Gap Analysis"]
    end

    L0 --> L1
    L1 --> L2
    L2 --> L3
    L3 --> L4
    L4 --> L5
    L5 --> M7
    M7 -->|"connects to L1"| M8
    L1 -.->|"SBT 8 dimensions"| M7

    style L0 fill:#1E3A5F,color:#fff,stroke:#BAE6FD
    style L1 fill:#264D73,color:#fff,stroke:#BAE6FD
    style L2 fill:#2E6087,color:#fff,stroke:#BAE6FD
    style L3 fill:#36739B,color:#fff,stroke:#BAE6FD
    style L4 fill:#3E86AF,color:#fff,stroke:#BAE6FD
    style L5 fill:#4699C3,color:#fff,stroke:#BAE6FD
    style M7 fill:#0d47a1,color:#fff,stroke:#BAE6FD
    style M8 fill:#004d40,color:#fff,stroke:#BAE6FD
Loading

What Is This?

Organizational Schema Theory applies Test-Driven Development methodology to business design. Businesses are designed backward from desired customer experience through testable, version-controlled specifications where each operational layer validates the layer above it.

This toolkit is an 8-module AI prompt pipeline. Copy prompts into any capable LLM (Claude, GPT, Gemini) and produce a complete operational specification -- from customer experience goals down to sourcing requirements. No framework knowledge required to start.

Quick Start

Zero to complete business specification in 40 minutes:

  1. Copy the prompt from prompts/01_EXPERIENCE_DESIGN.md into your LLM
  2. Describe your business in plain language: "I run a neighborhood bakery. We want customers to feel welcomed, trust our ingredients, and leave with something they are proud to share."
  3. Get structured YAML: experience contracts, signal requirements, process specifications
  4. Cascade through Modules 2-6, pasting each output into the next prompt (~5 min each)
  5. Validate with Module 8 -- the LLM identifies gaps, orphans, and conflicts
  6. Iterate -- fix gaps by revisiting the relevant module

For common business types, start with a starter pack instead of describing from scratch.

For automated validation, feed YAML output into orgschema-validate.

The 8 Modules

# Module Grounded In Input Output
1 Experience Design Design Thinking, JTBD Business description (plain language) L0 experience contracts
2 Signal Architecture SBT 8 Dimensions Module 1 output L1 signal requirements
3 Process Contracts SIPOC, ISO 9001 Module 2 output L2 process contracts
4 Procedure Specification Toyota Standardized Work Module 3 output L3 procedures
5 Input Specification Bill of Materials, SCM Module 4 output L4 input specs
6 Sourcing Requirements ISO 9001 clause 8.4 Module 5 output L5 sourcing requirements
7 Signal Mapping SBT Signal Decomposition Modules 1-6 output Operations-to-brand signal map
8 Gap Analysis TDD, Lean Muda All module outputs Validation report

Each module has a prompt in prompts/ and a YAML template in templates/. See templates/FRAMEWORKS.md for alternative framework citations per module.

Industry Starter Packs

Pre-filled L0 experience contracts for common business types. Copy, modify, and feed into Module 2.

Starter Contracts Types Included
Food Service 12 5 experience, 4 constraint, 3 commitment
Retail 11 6 experience, 3 constraint, 2 commitment
SaaS (B2B) 10 5 experience, 3 constraint, 2 commitment
Professional Services 9 5 experience, 2 constraint, 2 commitment
Education 10 6 experience, 2 constraint, 2 commitment

Executor Profiles

Procedures (Module 4) support three executor types in a single file:

Profile Description When to Use
human (default) Human operator following step-by-step procedures Most businesses starting out
hybrid Human with automated assistance (sensors, auto-dosing) Scaling with quality control
automated Fully automated system (IoT, robotics) High-volume or precision-critical

The active profile is selected by a single field: active_executor: human. Change it to hybrid or automated to switch. Module 8 (Gap Analysis) can compare profiles and flag signals that change when you switch executor types.

Scope

This toolkit produces a complete operational specification: from customer experience goals (L0) through sourcing requirements (L5), with signal mapping and validation.

The following domains are adjacent but outside this toolkit's scope:

  • Financial modeling (pricing, P&L, unit economics)
  • Labor management (hiring, scheduling, compensation)
  • Marketing and customer acquisition
  • Legal and corporate structure
  • Capital expenditure and facility planning

These domains interact with operational specifications -- a sourcing requirement has cost implications, a procedure has staffing implications -- but specifying them requires domain-specific frameworks beyond this toolkit's purpose. The operational specification produced here serves as structured input to those planning processes.

Key Concepts

Concept Meaning
TDD cascade Six-level test hierarchy: L0 acceptance tests through L5 infrastructure
Contract What must be achieved (the test) -- executor-invariant
Procedure How it is achieved (the implementation) -- executor-specific
Quality gate Measurable parameter with min/max/target that a process must satisfy
Forkability Copy the test suite, rewrite the implementation for your context
Schema vs Data Schema (what to measure) is publishable; Data (values) is competitive moat
Backward traceability Every parameter traces to the L0 contract it serves

See docs/GLOSSARY.md for the complete terminology reference.

Repository Structure

orgschema-toolkit/
├── prompts/                    8 LLM-ready module prompts
│   ├── 01_EXPERIENCE_DESIGN.md
│   ├── 02_SIGNAL_ARCHITECTURE.md
│   ├── 03_PROCESS_CONTRACTS.md
│   ├── 04_PROCEDURE_SPECIFICATION.md
│   ├── 05_INPUT_SPECIFICATION.md
│   ├── 06_SOURCING_REQUIREMENTS.md
│   ├── 07_SIGNAL_MAPPING.md
│   ├── 08_GAP_ANALYSIS.md
│   └── README.md
├── templates/                  YAML output schemas (one per module)
│   ├── 01_experience_contracts.yaml
│   ├── 02_signal_requirements.yaml
│   ├── 03_process_contracts.yaml
│   ├── 04_procedures.yaml
│   ├── 05_input_specifications.yaml
│   ├── 06_sourcing_requirements.yaml
│   ├── 07_signal_map.yaml
│   ├── 08_validation_report.yaml
│   └── FRAMEWORKS.md
├── starters/                   Industry starter packs
│   ├── food_service.yaml
│   ├── retail.yaml
│   ├── saas.yaml
│   ├── professional_services.yaml
│   └── education.yaml
├── docs/
│   ├── METHODOLOGY.md
│   └── GLOSSARY.md
├── CITATION.cff
├── LICENSE
└── README.md

Related Projects

Project Description
orgschema-framework Python validator + JSON Schema (orgschema-validate CLI)
orgschema-demo Spectra Coffee reference implementation (25 YAML files, CI/CD)
sbt-framework Spectral Brand Theory toolkit (brand perception analysis)

Citation

@article{zharnikov2026ost,
  author  = {Zharnikov, Dmitry},
  title   = {The Organizational Schema Theory: Test-Driven Business Design},
  year    = {2026},
  journal = {Zenodo},
  doi     = {10.5281/zenodo.18946043},
  url     = {https://doi.org/10.5281/zenodo.18946043}
}

Author

Dmitry Zharnikov -- spectralbranding.com | orgschema.com

1 | Getting Started

This repository is a documentation + template + prompt toolkit. There is no software to install and no Python package to build — every artifact is a human- and LLM-readable Markdown or YAML file.

To use the toolkit:

  1. Clone or download the repository.
  2. Open the prompt file for the module you want (see prompts/).
  3. Paste the prompt into any capable LLM (Claude, GPT, Gemini), then describe your business.
  4. Save the LLM's structured YAML output and feed it into the next module.

Project anchor: a top-level .here file marks the repository root for relative-path resolution. No pyproject.toml ships because the repository contains no Python package; CI Python checks skip when pyproject.toml is absent.

2 | Project Layout

orgschema-toolkit/
├── prompts/                    8 LLM-ready module prompts
├── templates/                  YAML output schemas (one per module)
├── starters/                   Industry starter packs (5 sectors)
├── docs/                       METHODOLOGY.md + GLOSSARY.md
├── output/                     Standard split (empty here — toolkit has no pipeline)
│   ├── figures/
│   ├── tables/
│   └── logs/
├── CITATION.cff                Machine-readable citation
├── LICENSE                     MIT (code)
├── LICENSE-data                CC BY 4.0 (templates, prompts, docs)
├── reproduce.sh                Standard-conformant stub (no pipeline)
├── .here                       Project root anchor
└── README.md

See the Repository Structure section above for the per-file listing.

3 | Quick Start

The toolkit has no computational pipeline to reproduce; the standard-conformant orchestrator is a stub that announces that fact:

./reproduce.sh
# OST toolkit — docs and templates; nothing to reproduce.
# See https://github.com/spectralbranding/orgschema-papers for paper mirrors.

For the actual user workflow (LLM-driven business specification in ~40 minutes), see the Quick Start section above.

For paper computational pipelines, see the per-paper mirrors under orgschema-papers.

4 | Dependencies

None. The toolkit ships only Markdown prompts, YAML templates, and documentation. A capable LLM (Claude, GPT, Gemini) is the only runtime requirement, and it is operated by the user outside this repository.

The companion validator and reference implementation live in separate repositories:

5 | Script Map

Not applicable. This repository contains no executable scripts beyond reproduce.sh (a no-op stub). The toolkit's primary artifacts are prompts and templates, mapped per module in The 8 Modules table above.

6 | Citation

@misc{zharnikov2026osttoolkit,
  author  = {Zharnikov, Dmitry},
  title   = {Organizational Schema Theory: AI Prompt Toolkit},
  year    = {2026},
  url     = {https://github.com/spectralbranding/orgschema-toolkit}
}

@article{zharnikov2026ost,
  author  = {Zharnikov, Dmitry},
  title   = {The Organizational Schema Theory: Test-Driven Business Design},
  year    = {2026},
  journal = {Zenodo},
  doi     = {10.5281/zenodo.18946043},
  url     = {https://doi.org/10.5281/zenodo.18946043}
}

Machine-readable citation data lives in CITATION.cff. GitHub renders the "Cite this repository" button from that file.

7 | Licence

Dual-licence discipline per the public-mirror standard:

  • Code — MIT, see LICENSE.
  • Data, templates, prompts, docs — Creative Commons Attribution 4.0 International (CC BY 4.0), see LICENSE-data.

Attribution: Dmitry Zharnikov (2026). Organizational Schema Theory: AI Prompt Toolkit. Concept DOI 10.5281/zenodo.18946043.

Last updated: 2026-05-29

About

8-module AI prompt toolkit for Organizational Schema Theory. Design businesses backward from customer experience.

orgschema.com/toolkit

Topics

yaml tdd operations-management ai-toolkit business-design llm-prompts orgschema

Resources

Readme

License

MIT, Unknown licenses found

Licenses found

MIT
LICENSE
Unknown
LICENSE-data
Activity
Custom properties

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages