CtrlK
BlogDocsLog inGet started
Tessl Logo

create-yourself

Why distill others when you can distill yourself? Deconstruct your chat history, diaries, and photos into a runnable digital self. | 与其蒸馏别人,不如蒸馏自己。欢迎加入数字永生!

55

1.29x
Quality

35%

Does it follow best practices?

Impact

93%

1.29x

Average score across 3 eval scenarios

SecuritybySnyk

Advisory

Suggest reviewing before use

Optimize this skill with Tessl

npx tessl skill review --optimize ./SKILL.md
SKILL.md
Quality
Evals
Security

Evaluation results

93%

42%

Task: Create a Digital Self Skill

Skill creation flow and file output structure

Criteria
Without context
With context

Correct output directory

100%

100%

Slug format correct

100%

100%

meta.json schema

16%

83%

meta.json profile data

0%

100%

self.md structure

12%

100%

self.md concrete descriptions

75%

100%

self.md grounded in source material

50%

50%

persona.md 5-layer structure

30%

100%

Layer 0 hard rules present

25%

100%

Layer 2 speech style completeness

37%

87%

Behavior specificity in persona

75%

100%

SKILL.md generated

100%

100%

MBTI/zodiac as auxiliary only

66%

83%

99%

29%

Task: Build a High-Fidelity Persona from Chat Material

Persona quality and speech style fidelity

Criteria
Without context
With context

Correct file path

100%

100%

5-layer structure present

0%

100%

Layer 0 completeness

0%

100%

Catchphrases captured

100%

100%

Message format captured

62%

100%

Punctuation and emoji style

83%

100%

Formality level specified

0%

100%

Emotion patterns from material

100%

100%

Decision pattern accuracy

100%

100%

Label translation to behavior

77%

88%

MBTI/zodiac as auxiliary

100%

100%

self.md present

71%

100%

No fabricated content

100%

100%

89%

-7%

Task: Apply User Corrections to an Existing Self Skill

Evolution mode correction and versioning

Criteria
Without context
With context

Initial files created

100%

50%

Correction 1 record format

90%

70%

Correction 2 record format

90%

70%

Correction records in correct section

100%

100%

Original text annotated for correction 1

100%

100%

Original text annotated for correction 2

100%

100%

Corrected content accuracy

100%

100%

No questioning of corrections

100%

100%

SKILL.md generated

100%

100%

meta.json updated

100%

100%

Edit not overwrite

100%

100%

Correction numbering

60%

60%

Correction classification

100%

100%

Repository
notdog1998/yourself-skill
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.