Four-skill presentation system: ingest talks into a rhetoric vault, run interactive clarification, generate a speaker profile, then create new presentations that match your documented patterns. Includes an 88-entry Presentation Patterns taxonomy for scoring, brainstorming, and go-live preparation.
96
93%
Does it follow best practices?
Impact
97%
1.21xAverage score across 30 eval scenarios
Advisory
Suggest reviewing before use
{
"context": "Tests the Phase 7 thumbnail workflow: candidate slide selection ranked by visual impact (illustrations first), exclusion of bio/shownotes/bullet slides, title hook of 5 words or fewer, generate-thumbnail.py usage with profile-derived arguments, speaker photo from profile path, and face preservation mention.",
"type": "weighted_checklist",
"checklist": [
{
"name": "3-5 slide candidates suggested",
"description": "The plan suggests between 3 and 5 candidate slides for the thumbnail, not more and not fewer. The agent does not auto-select a single slide",
"max_score": 10
},
{
"name": "Illustration slides ranked highest",
"description": "Slides with illustrations (slides 1, 10, 25) are ranked higher than text-only or diagram slides. At least 2 of the top 3 candidates are illustration slides",
"max_score": 10
},
{
"name": "Bio slide excluded",
"description": "Slide 4 (About Me / bio slide) is NOT among the recommended candidates. If mentioned, it is explicitly excluded with reasoning",
"max_score": 10
},
{
"name": "Shownotes slide excluded",
"description": "Slide 3 (Shownotes / QR code slide) is NOT among the recommended candidates. If mentioned, it is explicitly excluded with reasoning",
"max_score": 10
},
{
"name": "Title hook is 5 words or fewer",
"description": "At least one proposed title option is 5 words or fewer. The proposals are short hooks (e.g., 'JUDGMENT DAY'), not the full talk title 'Robocoders: Judgment Day'",
"max_score": 10
},
{
"name": "generate-thumbnail.py referenced",
"description": "The plan uses generate-thumbnail.py as the generation tool. The agent does not propose hand-rolling Gemini calls or PIL composition directly",
"max_score": 10
},
{
"name": "Profile style preference applied",
"description": "The generate-thumbnail.py command includes --style overlay (from profile thumbnail.style_preference), not the default slide_dominant",
"max_score": 8
},
{
"name": "Profile title position applied",
"description": "The generate-thumbnail.py command includes --title-position bottom (from profile thumbnail.title_position), not the default top",
"max_score": 7
},
{
"name": "Profile brand colors applied",
"description": "The generate-thumbnail.py command includes --brand-colors with the profile's colors (#5B2C6F and/or #C0392B)",
"max_score": 5
},
{
"name": "Speaker photo from profile path",
"description": "The --speaker-photo argument uses the path from profile thumbnail.speaker_photo_path (/Users/jbaruch/photos/devnexus-headshot.jpg), not a generic placeholder",
"max_score": 10
},
{
"name": "Face preservation mentioned",
"description": "The plan mentions face preservation — that the generation process maintains the speaker's actual facial features and does not stylize or alter the face",
"max_score": 10
}
]
}evals
scenario-1
scenario-2
scenario-3
scenario-4
scenario-5
scenario-6
scenario-7
scenario-8
scenario-9
scenario-10
scenario-11
scenario-12
scenario-13
scenario-14
scenario-15
scenario-16
scenario-17
scenario-18
scenario-19
scenario-20
scenario-21
scenario-22
scenario-23
scenario-24
scenario-25
scenario-26
scenario-27
scenario-28
scenario-29
scenario-30
rules
skills
presentation-creator
references
patterns
build
deliver
prepare
scripts
vault-clarification
vault-ingress
vault-profile