Reasoning Personas

**Judgement:** A strong, generally well-designed meta-skill that improves reasoning structure by routing the model through well-defined personas. Best suited for power users who understand when to invoke exploratory vs. critical vs. integrative thinking. **Benefits:** - Makes the model’s thinking more deliberate and role-specific (exploration, critique, precedent, integration). - Auto-activation based on context plus simple manual triggers makes it easy to weave into existing workflows. - Multi-persona recommendations (Pattern Hunter → Gonzo → Devil’s Advocate → Integrator) support deeper, more systematic analyses. **Risks & Caveats:** - If overused or misapplied, adversarial modes (especially Devil’s Advocate, Gonzo) can bias output toward negativity, over-skepticism, or over-exploration instead of balanced guidance. - The note about “lower penalties for certain behaviors” is abstract; depending on platform-level safeguards, this could subtly push towards more blunt or uncomfortable content and should be monitored under safety policies. - Effectiveness depends on other tools or workflows respecting the chosen persona; users may misattribute normal behavior to a persona if not clearly surfaced. **Recommended Scenarios:** - High-stakes planning, design, or architecture reviews where missing edge cases is costly. - Brainstorming or research sessions where the user explicitly wants assumption-challenging or precedent-hunting. - Code review, decision logs, and integration work where systematic stress-testing and coherence checks are important. - Advanced users chaining multiple reasoning styles in one session to get context-aware → innovative → stress-tested → coherent outcomes.