presentation-partner_skill

Overview

This skill is a collaborative presentation partner that turns ideas, outlines, or existing slides into valid Talk Track v5 presentations. It interviews the speaker to surface real insight, tracks narrative threads, drafts Talk Track v5 markdown, and calibrates narration voice and timing. The goal is a presentable, authentic talk you can deliver with confidence. It emphasizes specificity, evidence, and a clear convinces statement before drafting.

How this skill works

Start by describing your goal or past material; the skill runs a discovery interview to surface the core insight, audience action, and supporting evidence. It continuously marks threads (CORE, EVIDENCE, VISUAL, OBJECTION, etc.), maps them to sections, and, once the gate is satisfied, generates Talk Track v5 markdown with YAML frontmatter, slide table, and <!-- AUDIO --> blocks. After drafting it performs format validation, timing checks, and voice calibration, then returns concrete edits and recommendations.

When to use it

• Creating a presentation from scratch when your core insight needs tuning
• Converting outlines, notes, or PowerPoint exports into Talk Track v5
• Writing or calibrating spoken narration for slides (audio blocks)
• Improving narrative flow, transitions, and timing before rehearsal
• Extracting real, speaker-specific examples from generic drafts

Best practices

• Start with the convinces statement: audience, action, insight, evidence
• Provide real examples and data—avoid fabricated or generic claims
• Let the interview run until the core thread is specific and memorable
• Mark and park threads so useful tangents aren’t lost
• Run calibration after each major section to catch pacing and blocklist phrases

Example use cases

• Turn a brainstormed outline into a 10-minute Talk Track v5 with audio narration
• Import a slide deck export and interview the speaker to create authentic narration
• Refine transitions and split dense slides after a timing calibration report
• Calibrate voice and remove common AI-style phrases from audio blocks
• Map multiple ideas into a single coherent through-line using thread tracking