howells/fiction

notes

Collect and process inline editing notes from chapters. Finds %%NOTE:%%, %%EDIT:%%, %%CHECK:%%, and %%CUT:%% markers.

First seen Jan 27, 2026

Installation

$ npx skills add howells/fiction --skill notes

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from howells/fiction · top by installs.

npx skills add howells/fiction

Browse all from howells/fiction

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 17
Default branch main
Open issues 0
Status Active

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 3,208 B
  • docs SUMMARY.md 130 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 19 installs

SKILL.md

Scan chapters for inline editing notes and present them for processing.

Note Format

While reading your manuscript, add inline notes using Obsidian's comment syntax:

%%NOTE: observation or question%%
%%EDIT: specific change needed%%
%%CUT: text that should be removed%%
%%CHECK: continuity or fact to verify%%

These are invisible in Obsidian's reading view but visible when editing.

What This Does

  1. Greps all chapter files for %%.*%% patterns
  2. Extracts notes with chapter name and line number
  3. Groups by type (EDIT, NOTE, CHECK, CUT)
  4. Presents them for review or processing

Usage

/fiction:notes              # List all notes across chapters
/fiction:notes 5            # List notes in chapter 5 only
/fiction:notes --process    # Work through notes one by one
/fiction:notes --clear      # Remove all processed notes

If arguments provided: $ARGUMENTS

Output Format

## Editing Notes

### EDIT (4 items)
- **Chapter 2, line 3:** damp wool repeated
- **Chapter 7, line 45:** awkward transition
- **Chapter 12, line 88:** unclear antecedent
- **Chapter 12, line 92:** word echo "silent/silence"

### CHECK (2 items)
- **Chapter 2, line 21:** was this asked in ch1?
- **Chapter 9, line 156:** timeline - is this before or after the party?

### NOTE (1 item)
- **Chapter 15, line 200:** consider expanding this moment

### CUT (1 item)
- **Chapter 3, line 67:** this paragraph feels redundant

Processing Modes

List Mode (default)

Shows all notes grouped by type. Good for getting an overview.

Process Mode (--process)

Works through notes one at a time:

  1. Shows the note with surrounding context (5 lines before/after)
  2. Offers options: Fix it, Skip, Delete note
  3. For EDIT/CUT: suggests specific changes
  4. For CHECK: investigates and reports findings
  5. Removes the %%marker%% after processing

Clear Mode (--clear)

Removes all %%.*%% markers from chapters. Use after you've addressed everything.

When to Use

  • During read-throughs: add notes as you spot issues
  • Before revision: collect all notes to plan your editing session
  • After revision: clear processed notes

Workflow

  1. Read chapters, adding %%EDIT:%%, %%NOTE:%%, etc. as you go
  2. Run /fiction:notes to see everything collected
  3. Run /fiction:notes --process to work through them systematically
  4. Run /fiction:notes --clear when done

Implementation

When executing this command:

  1. Find all notes:
grep -rn '%%[A-Z]*:' chapters/*.md
  1. Parse each match: Extract type, content, file, line number
  1. Group by type: EDIT, NOTE, CHECK, CUT (and any others)
  1. For --process mode:

- Read context around each note - Present with options - Apply changes if requested - Remove the marker after processing

Related Commands

  • /fiction:edit — Automated line-level editing
  • /fiction:review — Story and craft feedback
  • /fiction:continuity — Cross-chapter consistency checks