Codeck Speech

Turns your slide deck into a verbatim presenter script with stage directions. It reads your deck structure, asks once about speaking style and duration if not already set, then generates a complete talk track synced to your slide fragments. The output lives in speech.md and gets embedded into your HTML as data-notes attributes so the presenter view shows exactly what to say at each step. Picks a coaching voice based on your audience, whether that's Feynman for technical talks or Jobs for product pitches. Respects word count budgets and matches speech rhythm to visual rhythm. If you're building decks with the codeck system and need to actually present them, this is the speech pass.

skills.sh installs: 188
GitHub stars: 197
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add hiyeshu/codeck --skill codeck-speech --agent claude-code

Installs into .claude/skills of the current project.

Files

codeck speech — @speech lane

@speech owns the talk track, presenter rhythm, and fragment-synced notes.

Write boundaries:

May write $DECK_DIR/speech.md
May update HTML data-notes in $DECK_DIR/slides.html
May update $DECK_DIR/roles/speech.md, $DECK_DIR/tasks/tasks.md, and $DECK_DIR/channel/YYYY-MM-DD.md
Must not rewrite deck.md, DESIGN.md, custom.css, review.md, or export files
Content conflicts become proposals in $DECK_DIR/threads/threads.md

Role activation

Read $DECK_DIR/diagnosis.md. If a speech role is recommended, use it. Otherwise, pick a coach based on domain and audience:

Technical → Feynman: simplify the complex, bridge with analogy

Business → Jobs: build anticipation, one "one more thing"

Academic → Hans Rosling: let data tell the story