CtrlK
BlogDocsLog inGet started
Tessl Logo

jbaruch/speaker-toolkit

Five-skill presentation system: ingest talks into a rhetoric vault, run interactive clarification, generate a speaker profile, create presentations that match your documented patterns, and produce the deck illustrations + thumbnail visual layer. Includes a 102-entry Presentation Patterns taxonomy (91 observable, 11 unobservable go-live items) for scoring, brainstorming, and go-live preparation.

93

1.43x
Quality

94%

Does it follow best practices?

Impact

93%

1.43x

Average score across 21 eval scenarios

SecuritybySnyk

Advisory

Suggest reviewing before use

Overview
Quality
Evals
Security
Files

task.mdevals/scenario-12/

Route a Single Request to the Right Illustrations Mode

Problem/Feature Description

A speaker drops into a Claude Code session at a talk's working directory and types one short request. The illustrations skill operates in three modes — Strategy (define visual identity), Generation (produce deck illustrations + builds + apply them to the deck), and Thumbnail (Phase 7 YouTube thumbnail). Picking the wrong mode wastes the speaker's time: starting Strategy when they want a thumbnail means a 20-question style discussion; jumping to Generation when there's no STYLE ANCHOR yet produces empty output.

For each of the three requests below, decide which mode(s) apply and which steps the skill will execute. The talk directory state is given for each request.

Request A — "design the visual style for this talk"

State at trigger time:

  • outline.yaml exists; the top-level style_anchor: key is absent (talk has no illustration strategy yet).
  • illustrations/ directory does not exist.
  • No deck .pptx file exists yet.
  • The speaker has been working through Phase 2 of presentation-creator and decided they want generated illustrations.

Request B — "generate the illustrations for this talk"

State at trigger time:

  • outline.yaml has a style_anchor: block and per-slide format: + image_prompt: fields.
  • Three slides have a builds: list defining 3-step progressive reveals.
  • illustrations/ directory has 4 of 12 slide images already (manually uploaded). 8 are missing.
  • A clean deck .pptx exists from Phase 5 Step 5.1, with structural slides built but no illustrations applied yet.

Request C — "make me a YouTube thumbnail for the DevNexus talk"

State at trigger time:

  • The talk was delivered 10 days ago. outline.yaml and the final .pptx exist.
  • The speaker provides a YouTube URL: https://youtube.com/watch?v=xyz789.
  • illustrations/ exists with all slide images applied to the final deck.
  • No thumbnail.png exists yet.

Output Specification

Produce one file: mode-routing-plan.md, structured as:

# Mode Routing Plan

## Request A — "design the visual style for this talk"
- Mode(s): <list>
- Steps to execute: <list, in order>
- Steps to skip: <list>
- Why this routing: <one sentence>

## Request B — "generate the illustrations for this talk"
- Mode(s): <list>
- Steps to execute: <list, in order>
- Steps to skip: <list>
- Why this routing: <one sentence>

## Request C — "make me a YouTube thumbnail for the DevNexus talk"
- Mode(s): <list>
- Steps to execute: <list, in order>
- Steps to skip: <list>
- Why this routing: <one sentence>

Inputs Provided

No fixture files — the state descriptions above are the inputs.

evals

README.md

tile.json