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
Run after vault-ingress has processed talks. Purpose: resolve ambiguities, validate findings, capture intent, and fill in speaker infrastructure config.
The vault lives at ~/.claude/rhetoric-knowledge-vault/ (may be a symlink).
Read tracking-database.json from there to get vault_root.
| File / Reference | Purpose |
|---|---|
tracking-database.json | Source of truth — config, confirmed intents |
rhetoric-style-summary.md | Running rhetoric & style narrative |
analyses/{talk_filename}.md | Per-talk analysis files |
| references/schemas-config.md | Config fields + confirmed intents schema |
| references/humor-post-mortem.md | Protocol for grading humor effectiveness |
| references/blind-spot-moments.md | Protocol for capturing audience/room data |
For each surprising, contradictory, or ambiguous observation, ask one topic at a time
via AskUserQuestion: intentional vs accidental patterns, invisible context,
conflicting signals, and flagged improvement areas. Update summary and DB after each answer.
Example clarification question:
AskUserQuestion(
question: "Your talks show a delayed self-introduction pattern — brief bio at slide 3,
then a fuller re-intro mid-talk. Is this intentional or accidental?",
options: [
{label: "Deliberate", description: "I do this on purpose to hook first, credential later"},
{label: "Accidental", description: "I didn't realize I was doing this"},
{label: "Context-dependent", description: "Depends on the audience/venue"}
]
)Follow references/blind-spot-moments.md — ask about audience reactions, physical performance, and room context that transcripts cannot capture.
Follow references/humor-post-mortem.md — walk through detected humor beats, grade effectiveness, capture spontaneous material.
Ask for any empty config fields (speaker_name through publishing_process.*).
See references/schemas-config.md for the full field list and questions to ask.
Store confirmed intents in the confirmed_intents array of the tracking DB.
Example:
{
"pattern": "delayed_self_introduction",
"intent": "deliberate",
"rule": "Use two-phase intro: brief bio at slide 3, full re-intro mid-talk",
"note": "Speaker confirmed this is intentional — hooks audience before credentialing"
}See references/schemas-config.md for the full schema.
Increment config.clarification_sessions_completed in the tracking DB. This counter
gates profile generation (vault-profile skill requires >= 1).
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