talk-stage1-extract
Stage 1 — Extract and structure source material into a talk summary. Auto-detects REX vs Concept type.
Install with Tessl CLI
npx tessl i github:FlorianBruniaux/claude-code-ultimate-guide --skill talk-stage1-extractEvaluation results
100%
85%Talk Summary: Infrastructure Migration Story
REX source detection and metric extraction
REX type detected
0%
100%
Correct file path
0%
100%
Metadata header block
50%
100%
Chronological narrative arc
12%
100%
Before→after metric format
50%
100%
Metric source attribution
0%
100%
No metric rounding without indication
33%
100%
Comprehensive metric extraction
37%
100%
Themes table format
0%
100%
Gaps checklist
0%
100%
Narrative Potential section
0%
100%
Recommendations section
0%
100%
Without context: $0.7492 · 2m 6s · 21 turns · 25 in / 6,320 out tokens
With context: $0.4713 · 2m 2s · 23 turns · 287 in / 6,267 out tokens
100%
83%Talk Summary: Opinion Essay on API Design Philosophy
Concept source detection and metrics-absent handling
Concept type detected
0%
100%
Correct file path
20%
100%
Metadata header block
37%
100%
Thematic narrative arc
20%
100%
No-metrics placeholder
0%
100%
No invented metrics
100%
100%
Themes table format
0%
100%
Gaps checklist
0%
100%
Narrative Potential section
0%
100%
Recommendations section
0%
100%
Research recommendation for Concept
0%
100%
Without context: $0.6480 · 2m 1s · 27 turns · 532 in / 6,308 out tokens
With context: $0.4233 · 1m 46s · 20 turns · 183 in / 5,387 out tokens
100%
61%Talk Summary: Observability Journey at a Scale-up
Hybrid source detection and contradicting metrics flagging
Hybrid type detected
0%
100%
Both components noted
60%
100%
Correct file path
0%
100%
Metadata header block
50%
100%
Contradicting metrics flagged
92%
100%
No single metric chosen
90%
100%
Before→after metric format
62%
100%
Metric source attribution
12%
100%
Themes table format
0%
100%
Gaps checklist
33%
100%
Narrative Potential section
0%
100%
Recommendations section
0%
100%
Without context: $0.5400 · 2m 12s · 25 turns · 2,407 in / 6,896 out tokens
With context: $0.4138 · 1m 59s · 18 turns · 24 in / 5,688 out tokens
- Evaluated
- Agent
- Claude Code
- Model
- Claude Sonnet 4.6
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.