SanctumOS Naming Rubric

SanctumOS is neuro-inspired, not neuro-identical. We mirror brain roles where it clarifies behavior (e.g., consolidation, gating, cleanup) and adopt computationally efficient mechanisms where software can outperform biology. Our invariants—provenance, append-only history, gated action selection—preserve the spirit of cognition while keeping the system debuggable and fast.

Naming Categories

Here's how we label our terminology as we develop it:

N1 – Neuro-faithful

Strong functional analogy (Hippocampus/Neocortex/Glymphatic; Basal Ganglia)

N2 – Neuro-inspired

Shape matches, mechanics diverge (Dream loop/DMN-ish)

C – Computational

No useful analogy; pure tech (UI, MCP plumbing, indexes)

SanctumOS models the human brain while remaining grounded in AI/tech conventions. To keep naming consistent and intuitive, every new component falls into one of four naming camps.

1. Global Modules → Technical Names

These are OS-level or infrastructure pieces that aren't specific to cognition.

Examples: MCP, UI, Kernel (Letta)

Naming rule: Plain tech terminology (short acronyms or descriptive engineering words)

Audience: Developers/sysadmins

Purpose: Clarity of function, not metaphor

2. Agent Modules → Neuroanatomical Names

These model functions of the human brain.

Examples: Broca (speech center), Thalamus (routing/refinement), Cerebellum (filter/reflex)

Naming rule: Single-word brain regions/networks, chosen to match function

Audience: AI/agent architects

Purpose: Reinforce the metaphor of SanctumOS as a cognitive architecture

Guideline: Choose macro-level regions (hippocampus, amygdala) rather than microanatomy, so non-specialists can still infer meaning

3. Letta/AI Extensions → Industry Terms

These are continuations of Letta or AI ecosystem concepts.

Examples: agents, tools, memory blocks, SDK

Naming rule: Preserve existing AI/ML vocabulary

Audience: Broader AI/ML community

Purpose: Leverage familiarity, avoid needless reinvention

4. Agents/Primes → Personal/Mythic Names

Agents themselves follow a distinct naming track.

Examples: Athena, Monday, Sentinel

Naming rule: Proper nouns (mythic, literary, or thematic), chosen for narrative identity

Audience: End users, collaborators

Purpose: Convey personality and individuality

Neuroanatomical Reference

| Brain Region | SanctumOS Component | Function | |--------------|-------------------|----------| | Thalamus | Input Preprocessing Hub | Routing, refinement, sensory relay | | Cerebellum | Real-time Filter | Reflex processing, motor control | | Basal Ganglia | Task Orchestrator | Action selection, habit formation | | Hippocampus | Memory Consolidation | Short-term to long-term memory | | Neocortex | Knowledge Integration | Complex reasoning, pattern recognition | | Glymphatic System | Memory Optimization | Waste clearance, memory pruning | | Broca's Area | Communication Layer | Language production, message processing |

Deeper Reasoning

The why of this rubric is as important as the rules:

Cognitive Fidelity

SanctumOS is not just an agent system — it's a cognitive OS. By naming agent modules after brain regions, we preserve the metaphor that "this is a brain," not just software stitched together. That makes the architecture intuitive, teachable, and extensible.

Division of Concerns

Tech names (MCP, UI, Kernel) = engineering plumbing
Neuro names = cognition, perception, memory
Industry names = AI standards everyone already recognizes
Mythic names = personality and identity

This prevents namespace collision and tells new developers immediately "what layer they're working on."

Scalability of Metaphor

The brain has many regions, but only ~10–15 high-level ones matter for modeling cognition. That's enough to cover SanctumOS's foreseeable roadmap without getting too esoteric. If a module feels shoehorned, that's a clue it may not be core cognition.

Recruitment & Onboarding

New contributors immediately see "Thalamus → routing," "Hippocampus → memory consolidation," and understand the metaphor without reading 50 pages of docs. It's sticky and pedagogical.

Future-proofing

As SanctumOS grows, we can always extend into new brain regions (limbic system, cortex layers, etc.) without breaking convention. If something doesn't map naturally to neuroanatomy, it probably belongs in the tech or industry bucket instead.

Implementation Guidelines

When to Use Each Category

Use Neuroanatomical Names When:

The component directly models a brain function
The analogy helps users understand behavior
The component is part of core cognitive architecture
The function maps clearly to a well-known brain region

Use Technical Names When:

The component is infrastructure or plumbing
No clear brain analogy exists
The component is OS-level or system-level
Clarity of function is more important than metaphor

Use Industry Terms When:

Extending existing AI/ML concepts
Maintaining compatibility with broader ecosystem
The component follows established patterns
Familiarity aids adoption

Use Mythic Names When:

Creating agent personalities
Naming user-facing entities
Building narrative identity
Conveying character and individuality

Naming Best Practices

Consistency: Stick to the category once chosen
Clarity: Names should immediately suggest function
Scalability: Leave room for future expansion
Accessibility: Avoid overly technical neuroanatomy
Memorability: Choose names that are easy to remember and pronounce

SanctumOS