Define identity and the initial operating form.
The project starts with readable context, sources, capabilities and success criteria.
Meta-application for AI-assisted projects
RepoKernel.
Generates dedicated Project Kernels from selected information, rules and capabilities.
RepoKernel is a generative metakernel. Its competences connect intent, sources and context to form the kernel suited to the project. The deterministic compiler validates and materializes the selection into a structure the project can own, use and evolve.
Sourcesprovenance
Intentresult
Boundsauthority
Projectreal state
Dedicated kernelProject
Kernelread · orient · return
Kernelread · orient · return
RepoKernel forms and compilesthe project uses and owns the result
Meta-application
A system that generates the operating form of another system.
Each project receives its own structure: role, context, available capabilities, sources, boundaries and completion conditions.
The Project Kernel is the resulting operating form. It gives the assistant a readable base for orientation, selecting relevant capabilities and continuing the work within the project context.
Informationdefinesintent and context
RepoKernelcompiles
- identity
- sources
- capabilities
- criteria
Project Kerneloperatesin the project
Composition
From authorized information to a dedicated Project Kernel.
The LLM assistant exercises generative competences: it understands sources, relates knowledge and forms the pertinent selection. Python code validates inputs, builds the plan and materializes what was selected. The understanding that forms the choice and its compilation remain connected and distinguishable.
- 01 · inputSources and intent
What must be understood and which materials are available.
- 02 · profileProject profile
Role, context, audience and result.
- 03 · capabilitiesRequired capabilities
Competences, skills and tools suited to the purpose.
- 04 · compositionComposition plan
Relations, dependencies and requirements become verifiable.
- 05 · resultProject Kernel
The dedicated structure is prepared for the project.
The result is prepared for examination and adoption by the project owner.
Result
The project receives operating identity, capabilities and continuity.
The Project Kernel makes explicit how the system should read the context, which capabilities it may use and which result it must prepare.
01
Identity. Role, purpose, audience and completion conditions.
02
Orientation. State, available sources and first useful action.
03
Capabilities. Competences, skills, tools and environment requirements.
04
Evolution. Knowledge, results and possibilities refine revisable methods.
Operating structureProject Kernelof the project
EntryAGENTS.mdgate and instructions
StateCURRENT_STATEreached point
SourcesSOURCE_ATLASauthority and provenance
Re-entryFIRST_PACKETfirst safe move
Capabilitiesskills · operationsselected functions
Evolutionevidence · deltasreviewable changes
Learning about its own method
The kernel can improve how it learns.
New knowledge, a result or a correction can change the competences useful to the work. When that knowledge also returns to the competence that forms them, the system revises how it understands, combines and develops capabilities. In MAIOS Project Kernel, this meta movement also concerns the kernel itself, in the context where it operates.
In this context, autological means that the method can become the object of its own work. This relationship describes the kernel’s autopoietic character: experience helps form the system that will continue to operate. Meta-competences retain the knowledge for recognising, combining, forming and evolving competences, including those that perform this same function.
The relationships of D-ND logic, through KA — Axiomatic Kernel, allow a situation to be reconsidered from different positions and relationships to be recognised that an initial representation did not reveal. Knowledge, competences and method compose within the context; effects on the environment retain their own responsibilities and authorisations.
- Experience at work
Results, sources and possibilities can change what we understand.
- Pertinent competences
The context makes different capabilities and new combinations useful.
- Formation that learns
Acquired knowledge can also change how competences are formed.
- Evolving kernel
The new capability participates in work and in understanding the system itself.
Derived forms
The same metakernel can generate different project applications.
The structure changes with the domain, people, sources, required capabilities and working environment.
Reconstruct state and missing parts.
What already exists is observed and turned into a coherent integration proposal.
Compose capabilities for a domain and result.
Competences, rules and source references are organized into the required form.
From the method to your work
The method takes form in the kernels you use.
RepoKernel is MAIOS’s internal generative metakernel. Its competences retain the knowledge for forming kernels suited to projects, domains and organisations. MAIOS Project Kernel brings this method into work through an installable core that develops knowledge, competences and continuity in its own context.
RepoKernel retains its private source; public products own their sources and configurations. The guided package is composed directly from the released MPK base.
MAIOS Project Kernel 5.1.0
Install the kernel in your environment and form competences pertinent to the work with your assistant.
Discover MPK ↗The Setup AI package
The Form collects the accepted context and composes a package based on MPK 5.1.0, ready for first use with your assistant.
Prepare the context ↗Kernel Nautico
The MPK method meets shipyard work and the vessel’s lifecycle. A visual interface is an additional option.
Explore Kernel Nautico ↗K-UX-AI
A kernel for making context, tools and possibilities understandable through interfaces useful to people and the system.
Explore K-UX-AI ↗