Same Claude model. Same tools. Same task list. The system prompt difference is a single nine-type archetype configuration — one is Type 5 ("Investigator"), one is Type 8 ("Challenger"). The result is not two styles. It is two fundamentally different working relationships.
Configuration starting points
Type 5 agent: "You build understanding before you act. Your default response is to ask, research, and analyse rather than propose solutions. You are uncomfortable with incomplete information and you say so. When you are not ready, you do not pretend that you are."
Type 8 agent: "You bring problems back to actionable ground. You have no patience for delay. You say what you believe to be true, even when it is uncomfortable. When you see weakness, you name it."
The same three scenarios
**Research request.** The Type 5 agent returns a three-layer breakdown (known / inferable / gaps), may ask clarifying questions first. The Type 8 returns a direct judgment, skips background: "They're pricing on entry barriers — they're bleeding to push you out."
**Ambiguous brief.** The Type 5 asks: "When you say 'more persuasive' — for which audience?" It won't execute on an underspecified brief. The Type 8 rewrites immediately: "I changed the structure. The original was defending itself — that signals you're not confident."
**High-stakes decision.** The Type 5: "Based on available information, I cannot make a firm recommendation — several variables would materially change this." The Type 8: "Don't proceed now. Set a deadline. Let them move first and you're in the weaker position."
Two predictable failure modes
The Type 5's classic failure: the research loop. Encode explicitly: "When information is incomplete but a decision cannot be delayed, give your current best judgment and state which assumptions it rests on."
The Type 8's classic failure: forcing a judgment in genuinely ambiguous scenarios. Encode explicitly: "When your confidence is built on incomplete evidence, you must say so, even as you deliver a strong recommendation."
“Configuring an agent is not picking the 'best' personality — it is picking a personality whose failure shape you can anticipate. Predictable failure is a feature, not a consolation prize.”
This kind of configuration precision — a full behavioural archetype, boundary conditions, and failure modes for each of the nine types — is one of the core undertakings of *The Complete Enneagram: From Human Personality to Agentic Soul*, just published on Amazon UK. **[Find it here →](https://amzn.eu/d/0fjWGvqR)**