Skip to content

semanticintent/project-phoenix

Repository files navigation

Project Phoenix

The Agentic Greenfield Pipeline for Legacy Modernization

DOI License: CC BY 4.0 GitHub

Don't translate legacy code. Extract the intent. Rebuild from zero.

"The code is not the asset. The business rules are the asset."


Abstract

Project Phoenix is a six-agent agentic pipeline that transforms legacy systems into modern, greenfield applications. Instead of translating old code line-by-line, Phoenix extracts the business logic -- the rules, workflows, and intent buried inside decades of legacy code -- and rebuilds from zero.

The name comes from the mythological bird (Bennu, ~1500 BCE) that burns itself to ash and is reborn. The old system dies. The knowledge lives.

Legacy modernization has been stuck for decades because the industry confused the implementation with the intent. AI changes the equation -- not by translating old code, but by understanding what the code was trying to do, then building it fresh.

Author: Michael Shatny ORCID: 0009-0006-2011-3258 Date: March 2026 License: CC BY 4.0 Website: phoenix.cormorantforaging.dev


The Old Model vs. The New Model

Traditional Migration Phoenix Pipeline
Approach Translate COBOL -> Java Extract intent -> Build fresh
Technical Debt Carried forward Zero
Timeline 18-36 months 3-6 months
Team 200+ consultants Small team + agent fleet
Cost $50M-$500M Orders of magnitude less
Failure Rate 70% stall or fail Validated at every gate

The Six-Agent Pipeline

Source Code --> [1] Extractor --> [2] Archaeologist --> [3] Synthesizer
                                                             |
               [6] Validator <-- [5] Builder <-- [4] Architect
Agent Role Output
1. Business Logic Extractor Source Code -> Business Requirements Rules catalog, workflow maps, edge cases
2. Interface Archaeologist UI/UX -> User Intent Model Screen inventory, journey maps, role matrix
3. Requirements Synthesizer Logic + Interface -> Unified Spec Single source of truth, gap analysis
4. Solution Architect Requirements -> Stack & Blueprint Tech stack, API contracts, build plan
5. Builder Fleet Blueprint -> Working Software Production codebase, CI/CD, documentation
6. Validator & Certifier New System <-> Original Requirements Regression proof, coverage report

Each agent is detailed in PIPELINE.md.


The Human Layer

Agents execute. Humans decide.

At every stage, the AI Software Lead validates outputs, resolves ambiguities, interviews stakeholders, and makes the judgment calls that no agent can:

  • Stakeholder interviews -- capturing tribal knowledge
  • Ambiguity resolution -- deciding what undocumented behavior means
  • Business sign-off -- confirming agent outputs match reality
  • Go/no-go decisions -- human gates between pipeline stages

The human layer is not optional. It is the difference between automation and transformation.

Details in HUMAN-LAYER.md.


The Phoenix Metaphor

The Phoenix myth (Bennu bird, Egypt, ~1500 BCE) is not about preservation. It is about deliberate destruction followed by intentional rebirth:

  1. The fire is voluntary -- the Phoenix chooses to burn. In software: the deliberate decision to stop maintaining the old system.
  2. The ashes carry the essence -- not the feathers or the nest, but the identity. In software: the business rules survive, the code does not.
  3. The rebirth is complete -- no hybrid Phoenix. A greenfield build, not a migration.

Full origin story in ORIGIN.md.


Cormorant Stack Integration

Phoenix is the terminal layer of the Cormorant Foraging stack:

     Cormorant Foraging
    (The Moon -- 3D Foundation)
   /         |         \
Sound     Space      Time
(Chirp)   (Perch)    (Wake)
   \         |         /
    \        |        /
     +-------+-------+
     |   Layer 1     |
     |   DRIFT       |    Measure the gap
     |  (See Gap)    |
     +-------+-------+
             |
     +-------+-------+
     |   Layer 2     |
     |   Fetch       |    Decide to act
     |  (Close Gap)  |
     +-------+-------+
             |
     +-------+-------+
     |   Layer inf   |
     |   GESA        |    Learn across episodes
     |  (Learn Why)  |
     +-------+-------+
             |
     +-------+-------+
     |   Phoenix     |
     |  (Rebirth)    |    Transform the system
     +---------------+

Phoenix is invoked when the DRIFT is so extreme that no incremental action can close it -- when the system itself is the problem. You don't optimize a burning building. You let it burn, preserve what matters, and rebuild.


Two Faces of Phoenix

Phoenix exists in two forms:

Face One: The Standalone Methodology

semanticintent.dev/phoenix

Phoenix as a self-contained legacy modernization methodology. No framework dependencies. No prerequisite knowledge. A CTO, consulting partner, or developer can read it cold and understand the entire proposition in five minutes.

Face Two: The Integrated Framework Layer

phoenix.cormorantforaging.dev

Phoenix as the terminal layer of the Cormorant Foraging stack. Full architectural integration showing how 3D sensing, DRIFT measurement, Fetch scoring, and GESA learning all feed into the Phoenix decision.

Same Phoenix. Same six agents. Same thesis. Two doors into the same room.


Economics

Traditional Migration Phoenix Pipeline
Timeline 18-36 months 3-6 months
Team size 200+ consultants Small team + agent fleet
Budget $50M-$500M Orders of magnitude less
Failure rate 70% stall or fail Validated at every gate
Technical debt Carried forward Zero
Architecture Modern wrapper around legacy complexity Built for the next decade

Details in ECONOMICS.md.


Repository Structure

project-phoenix/
+-- README.md                 # This file - framework overview
+-- PIPELINE.md               # The six agents in detail
+-- HUMAN-LAYER.md            # The human decision layer
+-- ORIGIN.md                 # The Phoenix myth and metaphor
+-- ECONOMICS.md              # Cost and timeline analysis
+-- CITATION.cff              # Machine-readable citation metadata
+-- CHANGELOG.md              # Version history
+-- CONTRIBUTING.md           # How to contribute
+-- ARCHIVING.md              # Zenodo archiving guide
+-- LICENSE                   # CC BY 4.0 license

Connection to Ecosystem

Project Phoenix is the transformation layer of the Semantic Intent ecosystem:

Semantic Intent (Philosophy)
  |
Cormorant Foraging (3D Framework) -- DOI: 10.5281/zenodo.18210081
  +-- PACE Pattern              -- DOI: 10.5281/zenodo.18049371
  +-- 6D Methodology            -- DOI: 10.5281/zenodo.18209946
  +-- DRIFT, Fetch (Action Layers)
  |
Project Phoenix (Transformation) <-- You are here

Related Projects:


Citation

If you use or reference Project Phoenix in your work, please cite:

APA Style

Shatny, M. (2026). Project Phoenix: The Agentic Greenfield Pipeline for Legacy Modernization. Zenodo. https://doi.org/10.5281/zenodo.18904231

BibTeX

@misc{shatny2026phoenix,
  author = {Shatny, Michael},
  title = {Project Phoenix: The Agentic Greenfield Pipeline for Legacy Modernization},
  year = {2026},
  publisher = {Zenodo},
  url = {https://github.com/semanticintent/project-phoenix},
  doi = {10.5281/zenodo.18904231},
  note = {ORCID: 0009-0006-2011-3258}
}

License

This work is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0).

You are free to:

  • Share -- copy and redistribute in any medium or format
  • Adapt -- remix, transform, and build upon for any purpose

Under the following terms:

  • Attribution -- You must give appropriate credit, provide a link to the license, and indicate if changes were made.

Author

Michael Shatny ORCID: 0009-0006-2011-3258


"You don't migrate a Phoenix. It burns and rises."

About

Six-agent agentic pipeline for legacy modernization via business logic extraction and greenfield rebuild

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Packages

 
 
 

Contributors