handdrawn-character-lock

Generates consistent hand-drawn character illustration series from a reference image using anchor locking.

Updated Aug 20, 2026
One-click install
npx skills add https://github.com/jkf87/character-deck-reviser --skill handdrawn-character-lock-jkf87
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: handdrawn-character-lock
Source: https://github.com/jkf87/character-deck-reviser/tree/main/skills/handdrawn-character-lock
Command: npx skills add https://github.com/jkf87/character-deck-reviser --skill handdrawn-character-lock-jkf87

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? When generating multiple hand-drawn illustrations of the same character across slides or scenes, AI image models drift in identity, style, and expression. This Skill locks the character's identity, line mode, palette, and expressions so every cut reads as the same person in the same art style. ## Core Features & Use Cases - Anchor-Based Identity Lock: Creates an approved anchor image from the user's character reference plus style-lock assets, then generates every scene from that same anchor in a star pattern instead of chained edits. - Style and Expression Control: Enforces clean or rough line modes, flat coloring, white backgrounds, and per-scene emotions with no default smile. - Plan Validation: Validates a JSON generation plan with a Python script before any image is created, checking paths, references, emotions, and object counts. - Use Case: A course creator needs six slide illustrations of the same cartoon character showing different emotions; the Skill produces a validated plan, an approved anchor, and six consistent cuts. ## Quick Start Use the handdrawn-character-lock skill to generate a five-scene hand-drawn illustration series from my character reference image at ./character.png in clean line mode.

Frequently Asked Questions about handdrawn-character-lock

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I keep an AI-generated character consistent across multiple images?

Create one approved anchor image from the original character reference, then generate every scene from that same anchor plus the original reference. Never use a previous scene's output as the next scene's reference, since chained edits cause identity drift.

How to generate hand-drawn style illustrations for presentation slides?

Define a line mode (clean for presentations, rough for social content), lock the character's hair, face, outfit, and colors in a plan file, then generate each scene with 1-4 supporting objects on a pure white background. Validate the plan with the included script before generating.

What is the difference between clean and rough line modes?

Clean mode uses smooth, even-weight closed ink contours suited for presentations and proposals, while rough mode uses jittery, broken pen strokes suited for SNS and carousels. The two modes must never be mixed within one series.

Why does my generated character look different in each scene?

Identity drift usually comes from chaining edits, where each new image references the previous output instead of the anchor. Regenerate the drifting cut from the original character reference plus the approved anchor, and request only one change per edit.

Can generated illustrations include Korean text labels?

Images contain no text by default since titles belong in the consuming PPTX as editable text. If Korean text is required, use only 2-6 character labels, visually verify the glyphs, and switch to numbers or English after two failed attempts.