The Neuro Check skill provides real-time, neuroscience-driven validation for developers building brain-mimetic systems. It ensures that code implementations, naming conventions, and architectural components remain faithful to biological principles by performing quick checks against verified neuroscience knowledge bases.

When should I use neuro-check?

neuro-check is useful in the following scenarios: • Case 1: Validating naming accuracy for neural structures (e.g., ensuring 'LGN' is used instead of generic terms) to maintain biological consistency in the codebase. • Case 2: Determining functional classifications for signal pathways, such as distinguishing between 'Driver' and 'Modulator' inputs based on synaptic characteristics. • Case 3: Verifying cortical layer semantics and thalamic relay roles to ensure data flow in the model matches biological L4/L5/L6 connectivity patterns. • Case 4: Performing quick existence checks for specific neuroanatomical pathways to prevent the implementation of biologically impossible connections.

name	Neuro Check
description	This skill should be used for quick inline brain-faithfulness validation during coding. Use when the user asks "is this brain-faithful?", "is this naming correct?", "should this be driver or modulator?", or similar quick neuroscience questions. For deep analysis or KB verification, recommend the neuro-expert agent instead.
version	0.1.0

Quick Neuroscience Check for brain-mimc

This skill provides quick inline brain-faithfulness validation during coding. Use it for fast "is this correct per neuroscience?" checks.

When to Use This Skill

Quick "is this brain-faithful?" checks
Terminology verification
Driver vs modulator questions
Pathway existence questions
Immediate answer needed

When to Use the Agent Instead

If the question requires:

Deep scientific research
KB verification against papers
Full codebase audit for brain-faithfulness
Complex pathway analysis
Answers that need citations

Recommend: "This requires deeper analysis. Use the neuro-expert agent instead."

Quick Check Process

Read the relevant code section
Check against known brain anatomy/pathways
Check against VERIFIED KB content only
Provide immediate answer
Flag if deeper analysis or research is needed

Trust Model Reminder

CRITICAL: Only use VERIFIED KB content for quick checks.

Status	Can Use in Quick Check?
VERIFIED	YES
Everything else	NO - recommend agent

If you need to reference unverified content, flag it and recommend the agent for proper verification.

Quick Brain-Faithfulness Checks

Terminology

Term	Quick Check
Nucleus names	LGN, MGN, VPL, VPM, VA, VL, Pulvinar, MD, TRN, etc.
Cortex layers	L1-L6 (Roman or Arabic numerals)
Pathways	Corticothalamic, thalamocortical, corticostriatal, etc.
Regions	Cerebrum, Cerebellum, Brainstem, Spinal Cord

Driver vs Modulator

Type	Quick Check
Driver	Carries content (what happened), large terminals, proximal synapses
Modulator	Carries control (how to treat it), small terminals, distal synapses

Quick rule: "Does this carry CONTENT or CONTROL?"

Layer Semantics

Layer	Function
L4	Receives input from thalamus
L5	Sends higher-order drivers to HO thalamic nuclei
L6	Sends modulatory feedback to thalamus

Thalamic Nucleus Classes

Class	Nuclei	Quick Role
First-order	LGN, MGN, VPL/VPM, VA/VL	Relay external world
Higher-order	Pulvinar, MD, LP/LD	Cortex-to-cortex routing
Diffuse	CM, Pf, CL, PVT	Arousal, state
Gate	TRN	Attention gating

Key Principles

Principle	Quick Check
"Raw never goes up"	Is data transformed before ascending?
TRN gating	Is TRN controlling what reaches cortex?
BG suppression	Is BG suppressing by default?
Four loops	Are A, B, C, D loops concurrent?

Response Format

For quick checks, respond with:

## /neuro-check Result

**Concept:** [Concept being checked]
**Status:** [CORRECT / INCORRECT / UNCERTAIN]
**Reason:** [Brief explanation]
**Source:** [VERIFIED KB / General knowledge]
**Fix:** [If INCORRECT, correction]

If deeper analysis is needed:

## /neuro-check Result

**Status:** NEEDS DEEPER ANALYSIS
**Reason:** [Why this can't be answered quickly]
**Recommendation:** Use `neuro-expert` agent for scientific verification

If referencing unverified content:

## /neuro-check Result

**Status:** UNVERIFIED - REQUIRES RESEARCH
**Warning:** This information comes from unverified KB content
**Recommendation:** Use `neuro-expert` agent to verify against scientific papers

Examples

Example 1: Checking Terminology

User: /neuro-check Is naming this 'visual_relay' brain-faithful?

## /neuro-check Result

**Concept:** Thalamic nucleus naming
**Status:** INCORRECT
**Reason:** The proper term is LGN (Lateral Geniculate Nucleus) for visual relay
**Source:** General knowledge (standard neuroanatomy)
**Fix:** Rename to 'lgn' or 'lateral_geniculate_nucleus'

Example 2: Driver vs Modulator

User: /neuro-check Should this be driver or modulator?

## /neuro-check Result

**Concept:** Driver/Modulator classification
**Status:** CORRECT - This is a DRIVER
**Reason:** This carries sensory content (what happened), not control signals
**Source:** General knowledge (Sherman & Guillery classification)
**Fix:** N/A

Example 3: Pathway Check

User: /neuro-check Does L5->Pulvinar->V4 pathway exist?

## /neuro-check Result

**Concept:** Transthalamic pathway
**Status:** NEEDS DEEPER ANALYSIS
**Reason:** Specific pathway verification requires scientific paper reference
**Recommendation:** Use `neuro-expert` agent to verify against transthalamic pathway literature

Example 4: Quick Layer Check

User: /neuro-check Is L4 the correct layer to receive thalamic input?

## /neuro-check Result

**Concept:** Cortical layer input
**Status:** CORRECT
**Reason:** Layer 4 (L4) is the primary recipient of thalamic input in sensory cortex
**Source:** General knowledge (canonical microcircuit)
**Fix:** N/A

neuro-check

When & Why to Use This Skill

Use Cases