Deep single-model analysis of a topic, question, or architectural decision.
Multi-model consensus workflow. Role-assigned models converge toward a shared recommendation through two rounds of evaluation and reconciliation.
Adversarial multi-model evaluation. Models argue for, against, and neutrally about a subject.
Multi-model panel review. Multiple models review independently, then findings are synthesized.
Full Forge QA checklist in Docker container. Use for release validation or comprehensive verification of all Forge features.
Pressure-test a claim, recommendation, or assumption. Defaults to skepticism.
Review design documents, specs, and technical writing for completeness and consistency.
Review code for conformance, correctness, and architecture alignment.