doubt-driven-development

Adversarial review is a fresh-context scrutiny process that extracts unstated assumptions, edge cases, and hidden dependencies from non-trivial decisions before artifacts are produced. It prevents confident-but-incorrect conclusions from propagating into production.

To use an artifact and contract for decision review, supply both elements to a fresh-context reviewer to enforce focused, testable analysis. The reviewer then captures adversarial findings, emphasizing issues and uncertainties rather than approving conclusions.

You need doubt-driven development for high-stakes scenarios like architecture changes, security-sensitive updates, or complex policy updates. It is required whenever context shifts could introduce unseen risks into non-trivial decisions before artifacts are produced.

Yes, adversarial review works explicitly for architecture changes, security-sensitive modifications, and complex policy updates. It enforces artifact and contract discipline to keep analysis focused and testable across these high-stakes contexts.

Doubt-driven development distinguishes itself from standard approvals by emphasizing the extraction of issues, uncertainties, and edge cases rather than simply approving conclusions. It uses adversarial prompts to surface hidden dependencies through fresh-context scrutiny.

A key limitation is its strict requirement for the user to supply an explicit artifact and contract before proceeding. Without these inputs, the fresh-context adversarial review cannot effectively extract unstated assumptions or evaluate non-trivial decisions.