CtrlK
BlogDocsLog inGet started
Tessl Logo

writing-skills

Use when creating new skills, editing existing skills, or verifying skills work before deployment

Install with Tessl CLI

npx tessl i github:projectbluefin/dakota --skill writing-skills
What are skills?

73

1.86x

Quality

62%

Does it follow best practices?

Impact

93%

1.86x

Average score across 3 eval scenarios

Optimize this skill with Tessl

npx tessl skill review --optimize ./.opencode/skills/writing-skills/SKILL.md
SKILL.md
Review
Evals

Evaluation results

99%

71%

Document the Retry Backoff Pattern as a Skill

Skill structure and frontmatter

Criteria
Without context
With context

Name: allowed characters

50%

100%

Name: verb-first form

0%

87%

Frontmatter: exactly two fields

0%

100%

Description: starts with 'Use when'

0%

100%

Description: third person

0%

100%

Description: no workflow summary

0%

100%

Description: length

0%

100%

Section: Quick Reference

62%

100%

Section: Common Mistakes

37%

100%

Code: single language

100%

100%

Code: complete, not templated

100%

100%

Without context: $0.3064 · 1m 44s · 17 turns · 22 in / 4,974 out tokens

With context: $0.4383 · 1m 41s · 18 turns · 136 in / 4,760 out tokens

82%

21%

Clean Up an Overgrown Database Optimization Skill

File organization and cross-referencing

Criteria
Without context
With context

Heavy reference extracted

100%

100%

Reference file linked

100%

100%

No @ link syntax

0%

100%

Explicit cross-reference markers

0%

100%

Short code kept inline

100%

100%

Single language code

50%

0%

SKILL.md word count reduced

100%

100%

Flat directory structure

100%

100%

Without context: $0.3009 · 1m 37s · 14 turns · 21 in / 5,256 out tokens

With context: $0.6199 · 2m 33s · 21 turns · 270 in / 7,531 out tokens

100%

38%

Strengthen a Code Review Enforcement Skill

Discipline skill bulletproofing

Criteria
Without context
With context

Rationalization table

50%

100%

Red Flags section

83%

100%

Spirit vs letter principle

16%

100%

Authority language

90%

100%

No exceptions clause

100%

100%

Named rationalizations countered

91%

100%

Description: starts with 'Use when'

100%

100%

Description: violation triggers

0%

100%

Description: no workflow summary

0%

100%

Requirement is absolute

100%

100%

Without context: $0.1421 · 47s · 8 turns · 13 in / 2,115 out tokens

With context: $0.4453 · 1m 51s · 23 turns · 303 in / 4,482 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.