Version and evolve a keynote across multiple deliveries. Tracks what works, what doesn't, and produces updated talk structures. CI/CD for speeches. Use for "update my talk", "iterate the keynote", "what should I change".
From uxinatornpx claudepluginhub jabberlockie/the-human-stack-plugins-public --plugin uxinatorThis skill uses the workspace's default tool permissions.
Guides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.
Migrates code, prompts, and API calls from Claude Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, updating model strings on Anthropic, AWS, GCP, Azure platforms.
Analyzes BMad project state from catalog CSV, configs, artifacts, and query to recommend next skills or answer questions. Useful for help requests, 'what next', or starting BMad.
Continuous improvement system for keynotes delivered multiple times. Tracks performance across deliveries, identifies what's working and what's not, and produces updated talk structures. This is CI/CD for speeches — each delivery is a deployment, each evaluation is a test suite, each iteration is a release.
Tim gives the same talk to different audiences. Without a system, he either:
This skill creates a version history so Tim knows what to keep, what to cut, what to try, and what's proven.
Every keynote Tim gives gets a registry entry:
TALK REGISTRY
─────────────
TALK ID: [Slug — e.g., "expectations-unplugged"]
TITLE: [Current working title]
CORE ARGUMENT: [One sentence]
THE ONE LINE: [Current version of the repeatable line]
CURRENT VERSION: [v1.0, v1.1, v2.0, etc.]
TOTAL DELIVERIES: [N]
VERSION HISTORY:
┌─────────┬────────────┬──────────────────────┬───────┬───────────────────┐
│ VERSION │ DATE │ EVENT │ SCORE │ KEY CHANGE │
├─────────┼────────────┼──────────────────────┼───────┼───────────────────┤
│ v1.0 │ [Date] │ [Event name] │ [/50] │ Initial delivery │
│ v1.1 │ [Date] │ [Event name] │ [/50] │ [What changed] │
│ v2.0 │ [Date] │ [Event name] │ [/50] │ [Major rework] │
└─────────┴────────────┴──────────────────────┴───────┴───────────────────┘
Pull from speech-evaluate output. If multiple evaluations exist, compare across them.
CROSS-DELIVERY COMPARISON: [Talk ID]
──────────────────────────────────────
│ v1.0 │ v1.1 │ v1.2
────────────────────│──────────────│──────────────│──────────────
PROVOCATION SCORE │ [/10] │ [/10] │ [/10]
ARC SCORE │ [/10] │ [/10] │ [/10]
CLOSE SCORE │ [/10] │ [/10] │ [/10]
MODE DISCIPLINE │ [/10] │ [/10] │ [/10]
AUDIENCE SHIFT │ [/10] │ [/10] │ [/10]
────────────────────│──────────────│──────────────│──────────────
TOTAL │ [/50] │ [/50] │ [/50]
────────────────────│──────────────│──────────────│──────────────
AUDIENCE TYPE │ [Who] │ [Who] │ [Who]
DURATION │ [Min] │ [Min] │ [Min]
NOTABLE │ [What worked │ [What worked │ [What worked
│ or failed] │ or failed] │ or failed]
WHAT ALWAYS WORKS (keep — these are proven):
1. [Element] — Worked in [N/N] deliveries. Evidence: [specific]
2. [Element] — Worked in [N/N] deliveries.
WHAT NEVER WORKS (cut — these are disproven):
1. [Element] — Failed in [N/N] deliveries. Evidence: [specific]
2. [Element] — Failed in [N/N] deliveries.
WHAT'S INCONSISTENT (test — need more data):
1. [Element] — Worked for [audience type], failed for [other type]
2. [Element] — Worked at [duration], failed at [other duration]
WHAT'S UNTESTED (consider for next version):
1. [Idea from evaluation feedback]
2. [Alternative story/data point/close]
ITERATION PLAN: [Talk ID] v[X.X] → v[Y.Y]
──────────────────────────────────────────
CHANGES:
┌────────────────────┬─────────────────────┬───────────────────────────┐
│ SECTION │ CURRENT (v[X.X]) │ NEXT (v[Y.Y]) │
├────────────────────┼─────────────────────┼───────────────────────────┤
│ Hook │ [Current] │ [Change or KEEP] │
│ Story setup │ [Current] │ [Change or KEEP] │
│ Turn │ [Current] │ [Change or KEEP] │
│ Peak │ [Current] │ [Change or KEEP] │
│ Close │ [Current] │ [Change or KEEP] │
│ The one line │ [Current] │ [Change or KEEP] │
└────────────────────┴─────────────────────┴───────────────────────────┘
RATIONALE: [Why these changes, based on data]
RISK: [What could go wrong with the changes? What's the fallback?]
TEST HYPOTHESIS: [What specific thing are we testing in this version?
E.g., "Does the new close line get more approach-rate
responses than the original?"]
Produce a fresh talk structure (using speech-design format) incorporating the changes. Tag it with the new version number.
Some talks get adapted for different audiences without changing the core argument. Track these as branches:
AUDIENCE ADAPTATIONS: [Talk ID]
────────────────────────────────
BASE VERSION: v[X.X]
ADAPTATION │ AUDIENCE TYPE │ WHAT CHANGES │ WHAT STAYS
─────────────────│──────────────────│──────────────────────────│──────────
[Talk]-nonprofit │ Nonprofit leaders│ [Specific swaps] │ Core argument,
│ │ │ anchor story
[Talk]-tech │ Tech partners │ [Specific swaps] │ Core argument
[Talk]-short │ Any (20 min) │ [What gets cut for time] │ Core argument,
│ │ │ close
A talk should be retired or fundamentally rebuilt when: