Agent Skill

Sepia skill

sepiaNanako0129/sepiaskills/sepia/SKILL.md

Make AI-generated writing read as human-written, in fiction and in professional prose. Repairs the narrative architecture of fiction and stories (based on StoryScope, arXiv:2604.03136); routes professional text through domain rules for release notes, announcements, PR and issue replies, code-review comments, incident postmortems, tickets, work orders, technical articles, blog posts, and long-form journalism. Four operations - write, review (diagnose AI tells without editing), refactor (minimal in-place edits), recreate (full rewrite). Use when asked to humanize, de-AI, unslop, or strip AI flavor from any text; when writing or revising any of these document types; or whenever output must not read as machine-written.

3.4kMIT

always on

155 tokname + description

once used

3,178 tokSKILL.md body

bundled files

20219 KB

repo stars

3.4kGitHub

Install the Sepia skill

Skills CLI (Claude Code, Codex, Cursor and others)

npx skills add https://github.com/Nanako0129/sepia --skill sepia

Claude Code, by hand (all projects)

git clone --depth 1 https://github.com/Nanako0129/sepia.git /tmp/sepia
mkdir -p ~/.claude/skills
cp -R /tmp/sepia/skills/sepia ~/.claude/skills/sepia

For one project only, copy the folder into .claude/skills/ in that project instead. The agent picks it up when a task matches its description.

What it costs your context

Agent Skills load in three steps. Tokens counted with the o200k tokenizer; your model's count will differ a little.

When	What loads	Cost
Every session	Name and description	155 tok
When the agent uses it	The SKILL.md instructions	3,178 tok
Only if opened	20 bundled files (219 KB)	varies

Until it is used, sepia costs about 8.8× less than the median MCP server we measured (1,363 tokens on every turn). Installing all 6 skills from Nanako0129/sepia would load 276 tokens of descriptions. Skills vs MCP study

Checks

Our checks found nothing to flag: it follows the Agent Skills specification's naming and size guidelines and doesn't pipe downloads into a shell.

Checked 2026-10-22 · method

What's inside

The headings of its SKILL.md. Read the full file on GitHub.

Sepia — de-AI writing
Security boundary
Routing
Operations
Fiction workflows
Calibration — the rule that governs all rules
Hard guardrails

More skills from Nanako0129/sepia

All skills in this repo

sepia-hemingwayNanako0129/sepia
Use when a user asks to write or revise fiction in the Hemingway manner, or asks for strong de-AI on a story; applies Sepia's built-in Hemingway voice profile.
3.4k43 tok idle · 355 in use
sepia-recreateNanako0129/sepia
Use when a user explicitly requests Sepia recreate for a full rewrite.
3.4k19 tok idle · 192 in use
sepia-refactorNanako0129/sepia
Use when a user explicitly requests Sepia refactor for minimal in-place prose revision.
3.4k22 tok idle · 193 in use
sepia-reviewNanako0129/sepia
Use when a user explicitly requests Sepia review to diagnose prose without editing.
3.4k19 tok idle · 192 in use
sepia-writeNanako0129/sepia
Use when a user explicitly requests Sepia write to create new prose.
3.4k18 tok idle · 192 in use