bmad-advanced-elicitation

To critique LLM-generated architectural decisions, you can apply structured critical thinking and elicitation methods like Socratic questioning and pre-mortems to pressure-test designs and identify hidden assumptions before implementation.

Structured elicitation is a critical review process that applies methodological techniques such as first-principles analysis to recent AI outputs. It forces a rigorous evaluation to refine unverified or superficial LLM-generated content before finalizing.

Yes, you can customize elicitation workflows by defining persistent preferences and custom methods via TOML configuration. This allows teams to tailor the iterative critique process to specific software engineering and planning needs.

Perform a red-team analysis on AI plans by invoking an advanced elicitation skill that applies a configurable catalog of critical thinking methods to identify potential failure points and hidden assumptions in recent conversation outputs.

Pytest is listed as a dependency for the elicitation and critique workflows. You will need it installed in your environment to run the scripts and components that perform the iterative, user-guided improvements on generated work.

Automated critique requires user-guided interaction for iterative improvements and does not automatically finalize work. It operates on recent conversation outputs, meaning it cannot evaluate external context or design decisions outside its immediate input scope.