Capability Clusters¶
G6Solver's 270 components are organized into 16 capability clusters -- functional groupings that reflect how components collaborate to deliver end-to-end cognitive capabilities. Nine are surfaced publicly through the REST
/clustersendpoint; the remaining seven are internal groupings. The table below does not yet list all sixteen.
Why Clusters?¶
Individual components are small, focused bricks. Clusters describe how those bricks compose into subsystems. A cluster is not a deployment unit -- it is a conceptual boundary that guides discovery, testing, and architectural reasoning.
Cluster Map¶
| Cluster | Components | Key Capabilities |
|---|---|---|
| Core Infrastructure | 19 | Result types, config, LLM routing, deployment, billing |
| Safety & Alignment | 10 | CSF verification, alignment specs, prompt governance |
| Formal Verification | 2 | SAT/SMT solving, CEGIS synthesis (405 formal_methods tools) |
| Context & Retrieval | 16 | RAG, ColBERT, Elastic, web scraping, document parsing |
| Goal & Planning | 6 | Goal decomposition, navigation, orchestration (325+ tools) |
| ML & Optimisation | 11 | Sklearn, Bayesian, genetic algorithms, deep learning |
| Data Processing | 3 | Pandas analytics, structured extraction, task queues |
| Code Intelligence | 8 | Metaprogramming, self-healing, debugging, security |
| Agents & LLM | 9 | Claude, OpenAI, LangChain, AutoGen, expert systems |
| Knowledge & Grounding | 3 | Domain knowledge, affordance KB, adaptive memory |
| Physical AI | 8 | Motor control, embodiment, physics, 3D meshes |
| Creative & Media | 11 | Audio, video, image, generative art, visualisation |
| Multimodal | 2 | Cross-modal reasoning, transformer integration |
| Experience & Autonomy | 9 | Experience loops, cognitive architectures, autonomy |
| Job Agents | 39 | 39 domain-specific agents on a shared framework |
Inter-Cluster Dependencies¶
graph TD
CORE[Core Infrastructure]
SAFE[Safety & Alignment]
FORMAL[Formal Verification]
CTX[Context & Retrieval]
GOAL[Goal & Planning]
ML[ML & Optimisation]
DATA[Data Processing]
CODE[Code Intelligence]
AGENT[Agents & LLM]
KNOW[Knowledge & Grounding]
PHYS[Physical AI]
CREAT[Creative & Media]
MULTI[Multimodal]
EXP[Experience & Autonomy]
JOBS[Job Agents]
CORE --> SAFE
CORE --> FORMAL
CORE --> CTX
CORE --> GOAL
CORE --> ML
CORE --> DATA
CORE --> CODE
CORE --> AGENT
SAFE --> FORMAL
SAFE --> GOAL
CTX --> GOAL
CTX --> KNOW
ML --> DATA
AGENT --> GOAL
AGENT --> CODE
KNOW --> CTX
EXP --> AGENT
EXP --> GOAL
JOBS --> CORE
JOBS --> AGENT
JOBS --> CTX
MULTI --> CTX
MULTI --> CREAT
PHYS --> ML Reading the diagram
An arrow from A to B means "cluster A provides foundations used by cluster B." Core Infrastructure sits at the root -- every other cluster depends on Result[T], AIBlock, and Settings.