| name | present |
| description | Structures a presentation narrative — the story arc, a slide-by-slide outline, and speaker notes — from results or receipts. Use to shape a talk or deck, or /present. Not for prose reports (report) or rendering visuals (asset). |
present
The narrative-structuring skill for reporting. Given results (receipts, a report, findings), it builds the
story of a presentation: the arc (setup → tension → resolution → ask), a slide-by-slide outline with one
message per slide, and speaker notes. It shapes the telling; it does not write the prose report (that is
report) or render the visuals (that is asset, which turns a slide spec into pixels on the DESIGN tokens).
It is read-only and advisory: it writes a PRESENTATION.md (the outline + notes), never product code or
rendered slides. The output is judged by Review — is there one clear message per slide, does the arc land
the ask, and is it grounded in the real results rather than spin?
What this does / does not do
Does: identify the audience + the single takeaway + the ask; shape the story arc; outline slides with one
message each (title + the point + what visual/data it needs); write speaker notes; ground every claim in a
receipt/result; write PRESENTATION.md + a receipt. Does not: write the prose stakeholder report (that is
report), render or design the actual slides/graphics (that is asset), or invent results not in the
evidence.
When NOT to use
- The deliverable is a written report/document, not a talk — use
report (multi-tier stakeholder prose).
- The need is the rendered slides/graphics themselves — use
asset (token-driven visual production) on this
outline.
- There are no results yet to present — run the work +
archive first; a presentation with no evidence is
spin.
- The audience needs a decision doc, not a narrative — use
propose (recommendation + tradeoffs).
Inputs
- The results to present —
state/receipts/*, a report output, findings, or stated outcomes.
- The audience + the occasion + the single takeaway/ask (or elicit them).
How to do it
- Fix audience, takeaway, ask. Who is in the room, the one thing they should remember, and what you want
them to do. Everything serves these.
- Shape the arc. Setup (why this matters to them) → tension (the problem/gap) → resolution (what was
done / found) → ask (the decision or action). A deck without an arc is a list.
- Outline slides, one message each. Per slide: a title that states the point, the single message, and
the visual/data it needs (named for
asset/dataviz to render — not rendered here).
- Write speaker notes. What is said around each slide — the connective tissue the slide does not show.
- Ground every claim. Each point traces to a receipt/result; no number or outcome that the evidence does
not support. Presentation is emphasis, not invention.
- Write
PRESENTATION.md — audience/takeaway/ask + arc + slide outline (message + needed visual) +
speaker notes.
Output
- PRESENTATION.md at the product root (or
presentations/PRESENTATION.md if that dir exists) —
persistence: Ephemeral. Audience + takeaway + ask · story arc · slide-by-slide outline (one message +
the visual it needs) · speaker notes · evidence trace.
- receipt (
<workspace>/receipts/present-receipt-<ts>.json): base schema + audience, takeaway,
ask, slide_count, arc, presentation_path. persistence: Ephemeral.
Completion Criteria
Done when: PRESENTATION.md fixes the audience/takeaway/ask, shapes a story arc, outlines slides with
one grounded message each (naming the visual each needs), and includes speaker notes; a receipt records the
takeaway + slide count.
Not done if: slides carry multiple messages (a list, not a narrative); a claim is not grounded in the
results (spin); or it rendered/designed the actual slides instead of outlining them (that is asset).
Next step: asset renders the slides/graphics from this outline on the DESIGN tokens; report covers a
written prose version for those not in the room.
Related Skills
- report: multi-tier stakeholder prose from receipts.
present structures the spoken/deck narrative;
report writes the document.
- asset: renders the actual slides/graphics on the design tokens.
present produces the outline + one
visual-need per slide; asset makes them.
- propose: a decision doc with tradeoffs. Use it when the room needs a recommendation, not a narrative.