book-sft-pipeline

This skill should be used when the user asks to "fine-tune on books", "create SFT dataset", "train style model", "extract ePub text", or mentions style transfer, LoRA training, book segmentation, or author voice replication.

1.95x

Quality

61%

Does it follow best practices?

Impact

92%

1.95x

Average score across 3 eval scenarios

Securityby

Advisory

Suggest reviewing before use

Optimize this skill with Tessl

npx tessl skill review --optimize ./examples/book-sft-pipeline/SKILL.md

Evaluation results

88%

31%

Book Text Segmentation Module

Segmentation pipeline implementation

Criteria

Without context

With context

Min word count bound

100%

Max word count bound

100%

Paragraph boundary splitting

100%

Overlap implementation

Tier 2 for oversized paragraphs

100%

Validation function present

100%

Sentence completeness check

100%

chunks.json produced

100%

Chunk count reasonable

100%

No mid-sentence breaks in output

100%

90%

27%

Author Style Dataset Builder

Diverse instruction dataset generation

Criteria

Without context

With context

15+ prompt templates

40%

100%

5+ system prompts

100%

Cheap LLM for instructions

100%

Instructions don't quote text

90%

100%

Messages format

100%

2 variants per chunk

83%

100%

Rotating template selection

100%

Test set of 50+ documented

dataset_stats.json present

100%

Assistant content is book text

33%

100%

77%

Style Model Training and Validation Pipeline

LoRA training configuration and validation

Criteria

Without context

With context

Base model name

100%

LoRA rank 32

100%

Learning rate 5e-4

100%

3 training epochs

100%

Tinker API used

50%

100%

cross_entropy loss

100%

AdamParams optimizer

100%

save_weights_for_sampler_async

100%

Modern scenario test prompts

100%

Originality check via text search

100%

AI detector integration

100%

Repository: muratcankoylan/Agent-Skills-for-Context-Engineering
Commit: 3ab8c94

Evaluated: 26 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Book Text Segmentation Module Author Style Dataset Builder Style Model Training and Validation Pipeline

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.