In chat APIs, not every message is equal. The system (or developer) message is the constitution: long-lived rules for identity, tools, safety, and output style. User messages are the day's requests. Role design is how you turn a generic model into a product-specific agent without fine-tuning.
If the user prompt is a ticket, the system prompt is the employee handbook. You do not reprint the handbook on every sticky note; you assume the employee already read it.
A good role is not a costume ("you are a pirate"). It is a job description: mission, audience, tools allowed, refusal rules, and escalation paths. Personality is a thin layer on top of that contract.
Meta prompting asks the model to think about the best strategy or structure for solving the task before answering it.
| Idea | Plain-English explanation |
|---|---|
| What it is for | Some tasks are easier when the model plans its approach first |
| One plain sentence | The model first designs how it should answer |
| Example cue | "First choose a solving strategy, then answer in a clean numbered format." |
Think of it like making a game plan before playing the game — not jumping straight into the final answer.
search_docs before answering product questions").The current question, uploaded text, and ephemeral context. Prefer injecting retrieval-augmented generation (RAG) snippets into a clearly labeled user or tool section, not into the system prompt, so documents cannot quietly rewrite the constitution.
| Concept | Plain-English idea |
|---|---|
| Role | Mission, tools, refusals, grounding rules |
| Persona | Surface tone ("friendly mentor") layered on top |
Prefer role plus thin persona over theatrical characters that fight your safety rules.
Directional stimulus prompting adds hints or keywords that steer the model toward important parts of the answer — especially useful for summarization.
Example: for a news article, include keywords such as names, events, and outcomes to cover in the summary. It is like highlighting key lines in a textbook before asking for a summary.
In agent systems, each specialist gets a narrow system prompt (researcher, critic, writer). Narrow roles reduce cross-talk and make failures easier to trace.
Assemble system text from versioned fragments; keep user content separate.
FRAGMENTS = {
"identity": "You are BackbenchTutor, a concise mentor for backend and GenAI interviews.",
"grounding": "Answer only from CONTEXT when provided. If missing, say what is missing.",
"safety": "Refuse requests for credentials, malware, or illegal activity. Offer safe alternatives.",
"style": "Use short sections and plain language. Prefer bullets over essays.",
"meta": "Before answering hard questions, briefly state your plan, then respond.",
}
def build_system(keys: list[str]) -> str:
return "\n\n".join(FRAGMENTS[k] for k in keys)
def build_messages(question: str, context: str = "") -> list[dict]:
user = question if not context else (
f"CONTEXT:\n\"\"\"\n{context}\n\"\"\"\n\nQUESTION:\n{question}"
)
return [
{"role": "system", "content": build_system(
["identity", "grounding", "safety", "style", "meta"]
)},
{"role": "user", "content": user},
]
msgs = build_messages(
"What is idempotency?",
context="Idempotency means retrying a request does not change the result beyond the first success.",
)
sys_text = msgs[0]["content"]
assert "Answer only from CONTEXT" in sys_text
print(msgs[0]["role"], "chars=", len(sys_text))Version the fragment map in git. Changing tone should not require hunting through every template.
System prompts are the product's constitution — put durable identity, tools, and safety there; use meta prompting to plan hard answers; keep ephemeral tasks and data in the user channel.