Open any scenario, click Read / Listen — long-form watch, then Update Narrative Story (top right). Available to owners, admins, and researchers.
The rubric modal has four toggle groups.
Length
- Short — 5 body paragraphs of ~120 words each
- Medium — 6 body paragraphs of ~160 words each
- Long — 7 body paragraphs of ~200 words each (default)
- Epic — 9 body paragraphs of ~240 words each
Tone
- Measured — steady throughout, dominant note is calm observation (default)
- Tense — dramatic tension allowed, sentences tighten where stakes rise
- Urgent — audible urgency, brief sentences, concrete consequences (still never panicked)
Perspective
- History walkback (default) — Numen writes from late 2030, looking back at how the scenario actually unfolded. Past tense for events that happened; present tense for lessons that hold. Headlines read like real news from the months ahead.
- Present anticipation — Numen writes from today, looking forward. Conditional structures (would, may, is likely to).
Specificity
- High — name companies, name analysts, name policy actors, regulatory bodies, dollar figures, percentages
- Medium — restrained particulars, only the details that hold
What gets composed
Every rubric run yields: a story title, an intro, 5–9 fictitious newspaper-style headlines (the spine), the body paragraphs, signposts, implications, an optional pull quote, an optional "from the future" passage, and a Numen for {your org} section at the bottom with summary + 3–7 specific actions to take now.
How long?
Roughly 30–90 seconds. Numen uses Anthropic's Sonnet 4.5 with prompt caching, so re-runs land faster than first composes.
Transparency
The structure of the prompt is public — see the prompt skeleton link inside the rubric modal. We do not hide how Numen composes scenarios.