advanced-elicitation

Invoke the advanced-elicitation skill after content is generated. It presents 5 context-selected critique methods such as socratic questioning or pre-mortem analysis, applies your chosen method to the content, and iterates until you confirm the enhanced version.

The skill loads methods from a methods.csv registry grouped by category, including core, structural, and risk techniques like first principles, pre-mortem, and red teaming. You can list all methods with the 'a' command or reshuffle for 5 new options with 'r'.

Yes, party mode resolves agent personas from sibling skill directories containing an [agent] block in customize.toml, such as analyst, PM, UX designer, architect, and developer roles. These personas join multi-persona elicitation methods with project-level overrides applied.

Use structured elicitation when content carries high risk, complexity, or stakeholder impact and informal revision requests miss blind spots. Named methods like pre-mortem and red teaming enforce systematic perspectives that ad-hoc feedback typically skips.

No, it always asks for confirmation before applying changes. After each method runs, you choose whether to accept the proposed enhancements, discard them, or give other instructions, and the loop continues until you select 'x' to proceed.