talk-stage2-research
Stage 2 (REX mode only) — Git archaeology, changelog analysis, verified factual timeline. Skip automatically in Concept mode.
Install with Tessl CLI
npx tessl i github:FlorianBruniaux/claude-code-ultimate-guide --skill talk-stage2-researchEvaluation results
0%
-44%Talk Research Package: chalk Terminal Styling Library
Output format & file structure
Three files created
100%
0%
Global Metrics table present
62%
0%
Source column populated
40%
0%
Monthly Velocity table
0%
0%
Contributors table
62%
0%
Commands run documented
62%
0%
Patterns section in changelog
0%
0%
Verifiable metrics in changelog
12%
0%
Timeline month-by-month table
40%
0%
Cross-reference section in timeline
50%
0%
File footers present
0%
0%
Period and repo header in archaeology
100%
0%
Without context: $1.1936 · 4m 31s · 54 turns · 844 in / 16,298 out tokens
With context: $1.3593 · 4m 35s · 48 turns · 313 in / 16,297 out tokens
99%
21%Fact-Checking Research for Express.js Retrospective Talk
Cross-reference & conflict flagging
Cross-reference section exists
90%
100%
Summary claims addressed
100%
100%
Unverified dates labeled
83%
91%
Conflicts noted not silently merged
100%
100%
Metrics have explicit source
25%
100%
Verifiable CHANGELOG metrics
75%
100%
Three output files present
100%
100%
Read-only git commands only
100%
100%
Monthly Velocity table
0%
100%
Contributors table present
100%
100%
Patterns section in changelog
62%
100%
Without context: $1.0032 · 4m 13s · 43 turns · 40 in / 14,762 out tokens
With context: $1.4891 · 5m 39s · 48 turns · 44 in / 21,764 out tokens
99%
21%Sprint Retrospective Research: execa's v5 Rewrite Sprint
Granularity & quiet period completeness
Week-by-week granularity
100%
100%
Quiet weeks included
100%
100%
Quiet periods annotated
80%
100%
Rounding flagged
90%
100%
Three output files present
100%
100%
Period fully covered
100%
100%
Metrics have explicit sources
37%
100%
Contributors table present
100%
100%
Patterns section in changelog
12%
87%
Cross-reference section in timeline
0%
100%
Read-only git commands
100%
100%
Without context: $1.7802 · 6m 44s · 59 turns · 65 in / 20,401 out tokens
With context: $1.2291 · 4m 44s · 58 turns · 164 in / 17,642 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.