workshop-facilitation

Workshop Facilitation is a protocol skill that defines a canonical pattern for running interactive, multi-turn guided sessions. It solves a common problem in AI-assisted facilitation: sessions that dump too many questions at once, hide progress, or repeatedly re-ask things the user has already answered. Instead, it establishes a disciplined one-step-at-a-time flow with clear pacing, visible progress, and predictable handling of interruptions, so any PM workshop or advisory session stays on track and ends with actionable choices.

The skill centers on a set of behaviors: an opening heads-up that sets expectations, an entry-mode choice (Guided, Context dump, or Best guess), one targeted question per turn, and user-facing progress labels such as "Context Qx/8" and "Scoring Qx/5." It offers quick-select numbered response options with an "Other (specify)" escape hatch, parses flexible selections like "1 and 3" or "1,3" into multi-select choices, and reserves enumerated recommendations for genuine decision points rather than after every answer. It also credits inline context supplied at invocation as answers already given, handles meta questions and stop/pause requests gracefully, and closes with a summary plus an "Assumptions to Validate" list when best-guess mode was used. An accompanying example transcript demonstrates the inline-input rule.

It is aimed at product managers, facilitators, and skill authors who build interactive Claude/AI skills. Typical use cases include running positioning workshops, discovery sprint kickoffs, retrospectives, and scoring or maturity assessments, or serving as the shared facilitation backbone that other workshop and advisor-style skills reference.