Skip to content
SkillModel-led205 tokens in use

Sepia Refactor

A fixed entry to the Sepia writing skill that runs only the refactor operation, listing defects first and then fixing them in place with minimal edits.

3,912 stars on the repository MIT LicenseNanako0129/sepia · skills/sepia-refactor
npx skills add Nanako0129/sepia -g
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Sepia Refactor costs

21tokens idle
205tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Fair
  • Frontmatter15/15
  • Structure0/25
  • Examples0/15
  • Output5/15
  • Size15/15
  • Resources10/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity88
  • Maintenance100
  • Community52
  • Maturity16
  • Momentum40

Sepia Refactor is a thin wrapper that binds the Sepia skill to its refactor operation. It reads the canonical Sepia skill from the sibling folder, which makes a full defect list first and then fixes each item, deepest layer first, keeping the structure, voice and intent of your text.

It never switches to a full rewrite on its own. It needs the complete Sepia package installed beside it.

What You Get

  • A defect list followed by in-place fixes to your existing text
  • Edits kept small, with a deletion test on additions and a reversion test on replacements
  • Notes on what was deferred or protected and left alone

Common Use Cases

  • Cleaning the tells out of a story draft without rewriting it
  • Tightening a PR reply or release note while keeping the author’s voice
  • Revising named files only, with protected line ranges left untouched

What Makes It Unique

It keeps the existing text and changes as little as it can, where a humanizer pass usually paraphrases everything. The two-stage protocol, protected ranges and unattended mode come from the canonical skill, and this entry will not swap in a different operation because of what the text contains.

Under The Hood

This is a prompt-only skill of one short file. It tells the agent to read the sibling canonical Sepia skill and follow it with the refactor operation. The skill itself runs no commands and calls no services. Edits happen through whatever file access the agent already has, and only to the files you name.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers