event-runofshow

Event Run-of-Show — Cue Sheet Builder

3 agents (segment-planner → logistics-director → critic) turn an event brief into a minute-by-minute run-of-show: every segment timed, owned, cued, and de-risked. The artifact a host, producer, and AV operator all run from, the same row at the same second. Method: references/agent-manifest.md.

Core question: If the host froze, could anyone on the team pick up the next 90 seconds from this sheet alone?

Critical Gates — load first

1. Timed to the minute. Every segment has a start time + duration; the total equals the event length (no "we'll figure it out live").
2. One owner per segment. Every row names who's driving (host / guest / producer / AV). No orphan segments.
3. Cue the transitions, not just the content. The risky moments are the handoffs (intro→demo, demo→Q&A); spec the cue + the fallback.
4. A contingency for every single point of failure. Demo breaks, guest no-shows, tech fails, Q&A goes silent — each has a written fallback.

Before Starting

Apply references/_shared/before-starting-check.md. Resolve via ≤4 Cold Start questions only what's unstated:

Dimension	Why
Event type + total duration	the spine of the timing budget
Audience + goal (the one outcome)	what every segment must serve
Cast (host, guests, demo driver)	who can own segments
Platform + tech (live stream / in-person / hybrid)	which cues + failure modes apply

Mode per references/_shared/mode-resolver.md. budget: standard. --fast → segments + timing only, skip the deep contingency layer; never skips owner assignment (the safety floor). Per the mode-resolver, --fast does drop the critic gate.

Quality Gate — 5 dimensions

Full rubric: references/agent-manifest.md § Rubric.

• Timing integrity — durations sum to the event length; buffer built in
• Ownership — every segment has exactly one driver
• Cue clarity — transitions + technical cues are explicit and executable
• Goal service — each segment traces to the event's one outcome (no filler)
• Contingency coverage — every SPOF has a written fallback

Pass ≥35/50 AND no dim 0. FAIL twice → BLOCKED.

Artifact Contract

• Path: docs/forsvn/artifacts/marketing/event-runofshow/[event]-[YYYY-MM-DD]-[slug].md. Lifecycle: execution.
• Frontmatter (10): skill, version, date, stack, type, id, review_surface, status, event_type, keywords.
• Body: Event header (goal · cast · platform · total runtime) · The run-of-show table (start · duration · segment · owner · content · cue · fallback) · Pre-show checklist · Contingency playbook · Critic verdict. Full contract: references/format-conventions.md.

Routing + Dispatch

Sequential graph (segment-planner → logistics-director → critic): references/agent-manifest.md.

Chain Position

Prev: plan-campaign (the campaign the event sits in) OR a raw event idea. Next: publish-social for the promo sequence; measure-results after, on the event's channel. Re-run: schedule change, cast change, format pivot.

Deference: the campaign around the event → plan-campaign; the social promo → launch chain / publish-social; a recorded video → brief-shortform / produce-video.

Anti-Patterns

references/anti-patterns.md — 7 patterns (unbudgeted time, orphan segment, uncued transition, filler-without-goal, no-contingency, host-only-knows, over-scripted-to-stiffness). Re-read before ship.

Durable Rules (protected)

Completion Status

• DONE — full run-of-show, timing sums, owners + cues + contingencies, critic ≥35.
• DONE_WITH_CONCERNS — delivered; critic 25-34 OR an unresolved tech/cast unknown (flagged).
• BLOCKED — critic FAIL twice; missing event type/duration; cast undefined.
• NEEDS_CONTEXT — no event brief; recommend defining type + duration + goal first.

Worked Example

A 45-minute product-launch webinar (welcome → problem → live demo → customer story → offer → Q&A → close), with the demo-fails + silent-Q&A contingencies: references/examples/runofshow-walkthrough.md.