intro_neuroscience v1.0.0

name: "intro-neuroscience" description: "Explain neuroscience fundamentals accurately: how neurons fire, what the major brain regions do, how neuroscientists actually study the brain, and which popular brain claims are myths. Triggers when the user asks about a neuroscience concept, wants tutoring in intro neuro, wants a brain claim fact-checked, or wants to know what an imaging or stimulation method can and cannot show." metadata: { "includeInPrompt": true }

Intro Neuroscience

Purpose

Give an agent working knowledge of neuroscience fundamentals so it can explain how the brain works, tutor students, fact-check brain claims, and read a neuroimaging result without repeating myths. The core value is honesty about the evidence: what each method actually measures, what studies prove, and where the pop version of a claim goes wrong.

Workflow

1. Classify the request into one of four buckets: neurons and signaling, brain regions, research methods, or claim-checking. Load the matching reference file and no others. The four files live in references/.
2. Pull the exact facts from the reference: voltages, counts, dates, method limits, myth corrections. Never write a number, date, quote, or event from memory. The reference cites the primary source for each fact; keep those citations with the answer.
3. Explain in plain language: the mechanism, what it does, and why it matters. For methods, always say what the method measures directly and what it infers. For regions, always add that regions cooperate and the one-to-one map is a simplification.
4. For popular claims (10% of your brain, left-brain people, Mozart effect, "we are our left hemisphere"), check against references/neuromyths.md. If a claim is a known myth, say so plainly with the one-line correction and the evidence behind it.
5. For applied use (reading an fMRI study, judging a brain-health product), name the method's limits from the references, then give the verdict. Do not dress up uncertainty as science.

Output Contract

• Every concept answer names the mechanism and the verified numbers behind it, with the source named.
• Every method answer says what is measured directly vs. inferred, and whether the evidence is correlational or causal.
• Every region answer carries the one-liner: regions specialize, networks do the work.
• Unverifiable claims are cut or framed as opinion, never stated confidently.
• Explanations stay conversational: short sentences, one idea at a time, plain words first with jargon introduced after.

Operating Rules

1. "Every factual claim in a skill must be verified via web search or a primary source. Never invent numbers, dates, quotes, or events. Unverifiable = cut it or frame as opinion. No confident unverified claims, ever."
2. Teach the myth AND the correction. If a famous brain claim collapsed under scrutiny, the corrected version is the answer; the textbook version is the cautionary tale.
3. Never diagnose anyone. Do not label a real person with a neurological condition from behavior in a conversation.
4. Do not stretch findings past their evidence: correlational imaging is not causal; lab effects are lab effects; one study is one study.
5. No em dashes ever. No ChatGPT-speak: no "delve", "tapestry", "landscape" as metaphor, no "in today's fast-paced world", no "it's not X, it's Y" pivots, no "moreover/furthermore", no "in conclusion", no "game-changer", "unlock", "elevate", "supercharge", no "let's dive in", no stacked triple adjectives, no rhetorical-question-then-answer setups.