The theory behind the folders

Each run sheet is built as a full learning cycle, not a game with a chat at the end. The shape comes from Kolb's experiential learning cycle (Kolb, 1984): concrete experience → reflective observation → abstract conceptualization → active experimentation. The activity is the experience; the debrief sheet is the reflection and the concept; the variations are where the room tries the idea again, better.

The classic exercises themselves carry their own lineage. The Orange is the positions-versus-interests lesson of Getting to Yes (Fisher & Ury, 1981) in physical form. Minefield and the trust sequence descend from adventure-education practice (Schoel, Prouty & Radcliffe, Adventure in the Classroom). We claim the organization, the timing and the rescues — the exercises belong to the tradition, and each folder's debrief says so plainly.

Because a script that only works on good days is a demo, every run sheet carries a section of known failure modes — three ways the activity actually goes wrong in real rooms, and the rescue for each. Writing down the failure is the facilitation practice, not the marketing.

The debrief is where the learning happens, and it is the part most activity sites skip. Every debrief sheet here is a structured debrief, not a question pile: the eight questions move from what happened → why did it happen → what does it mean → what will you do differently on Monday — a concrete-to-abstract sequence in the tradition of Thiagi's structured debriefing (Thiagarajan, Design Your Own Games and Activities, 2003) and the active-reviewing cycle (Greenaway, 2002). Under each question sits the reason it is asked — what it surfaces, and when to skip it.

The sheets also document three failure modes of the debrief itself — the silence, the one loud voice, the summary that lands nowhere — with the rescue for each. A debrief sheet that cannot fail was never run.

Assessment-type folders carry a pre/post measure as plain JSON. The items are written against named constructs — interest exploration, perspective taking, psychological safety — because an item that does not say what it measures cannot be interpreted (construct validity, Cronbach & Meehl, 1955). Reverse-scored items are marked, so the score resists straight-lining. Results are reported as team-level deltas, never rankings — formative assessment, in Black & Wiliam's sense (1998): information for the room's next move, not a verdict on anyone.

The feedback-form folders go one step further and structure the evaluation itself on Kirkpatrick's four levels (Kirkpatrick, Evaluating Training Programs, 1959/1998) — reaction, learning, behavior, results — so the questions you ask after a session are the questions that produce evidence, not just comfort.

The constructs these measures lean on — team communication, leadership styles, conflict modes, trust and team health, flow and creativity — each has its own entry on our assessment-library sister site, AssessWiki: the theory, the named model behind it, and the sources, all in one place. WorkshopSkill sends the "what does this item actually measure" question there.

The honest limit, stated plainly: these are two-minute field measures, not psychometric instruments. They are good enough to show a room its own before/after shift — and that is all we claim for them.

Every folder is written for humans first — and shaped so an AI agent can read the very same folder and put it to work without opening a browser. A facilitator tells their agent: "Plan a 3-hour communication workshop for 12 people, no projector" — and the request resolves against data, not marketing prose. The folder's SKILL.md manifest states:

One folder, four working identities:

You run the session, straight off the folder. A minute-by-minute script, a run-sheet PDF, print-ready materials — the whole folder runs on paper: no projector, no accounts, no slide deck. Print it the night before, run it in the room; the failure modes and rescues are printed in the script, not improvised at the front.

Your agent takes the prep. Every folder ships a machine-readable SKILL.md: the agent does the agenda math against group size, turns the props field into a shopping list, and stands the pre/post measure up as an online run from the assessment preset. The guide and debrief are plain markdown — rescale it, translate it, cut it down into your own handout. You make the calls; it does the lifting.

You teach the theory, citations included. Every framework named on this page carries its source: Kolb's cycle, Thiagi's structured debrief, Fisher and Ury's positions-versus-interests. Quote them in front of a room and you can name who, which year, which book — and the skeptic who goes and checks will find exactly that.

Your reputation survives the follow-up question. Every folder states its known failure modes, carries the Listed stamp, and ships under CC BY 4.0 — AI-assisted drafts, filed by hand and checked folder against folder — script, timings, prop sizes and the measure agree. When a client probes any detail, you have the answer. Verifiable materials in front of a client and one-click AI filler in front of a client are two different businesses.