by machaomc
Polish editable canvas strokes with safe patches, validation, handwriting normalization, and host integration contracts.
# Add to your Claude Code skills
git clone https://github.com/machaomc/refine-canvas-strokesGuides for using ai agents skills like refine-canvas-strokes.
refine-canvas-strokes is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by machaomc. Polish editable canvas strokes with safe patches, validation, handwriting normalization, and host integration contracts. It has 50 GitHub stars.
refine-canvas-strokes's catalog security scan is still queued. You can run an instant dependency and prompt-injection check now with the "Scan for vulnerabilities" button above.
Clone the repository with "git clone https://github.com/machaomc/refine-canvas-strokes" and add it to your Claude Code skills directory (see the Installation section above).
refine-canvas-strokes is primarily written in Python. It is open-source under machaomc on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh refine-canvas-strokes against similar tools.
No comments yet. Be the first to share your thoughts!
Unlocks once the catalog security scan passes (runs nightly).
The deep catalog scan for this skill is still queued. Run an instant dependency check now instead.
Refine Canvas Strokes is an open Agent Skill and Codex plugin for designing, validating, and integrating safe refinement of editable 2D canvas strokes.
It helps whiteboards, note apps, drawing tools, annotation systems, and pen-input SDKs improve freehand strokes, grouped handwriting, geometric shapes, and constrained diagrams without treating a rendered bitmap as editable source data.

The bundled scripts execute cleanup, handwriting normalization, validation, metrics, and comparison. Primitive fitting beyond the baseline, semantic flow reconstruction, OCR, and raster generation require a host implementation or optional external engine.
Add this GitHub repository as a Codex marketplace:
codex plugin marketplace add machaomc/refine-canvas-strokes
Start Codex, enter /plugins, select the refine-canvas-strokes-marketplace source, and install Refine Canvas Strokes. Start a new session after installation.
Install only the portable skill with the cross-agent skills CLI:
npx skills add https://github.com/machaomc/refine-canvas-strokes/tree/main/plugins/refine-canvas-strokes/skills/refine-canvas-strokes
Copy or link the skill directory into a supported project location such as .agents/skills/refine-canvas-strokes:
plugins/refine-canvas-strokes/skills/refine-canvas-strokes/
From the repository root:
python3 scripts/check_release.py
python3 scripts/build_release.py
Run a bundled example directly:
cd plugins/refine-canvas-strokes/skills/refine-canvas-strokes
python3 scripts/refine_strokes.py references/example-input.json --output /tmp/refinement-patch.json
python3 scripts/validate_contract.py patch /tmp/refinement-patch.json --input references/example-input.json
python3 scripts/compare_stroke_patch.py references/example-input.json /tmp/refinement-patch.json --fail-on-violation
Python 3.10 or later is required. Core scripts use only the Python standard library. extract_font_profile.py optionally requires FontTools.
| Automatic handwriting | Shape fitting guidance |
|---|---|
![]() |
![]() |
Editable source SVGs and additional comparison images are included under the skill's assets/examples/ directory. The handwriting and font examples use synthetic input and a synthetic font profile.
Read Privacy, Security, and Terms before production integration.
.agents/plugins/marketplace.json Codex repository marketplace
plugins/refine-canvas-strokes/.codex-plugin/ Codex plugin manifest
plugins/refine-canvas-strokes/skills/ Portable Agent Skill
review/ Public submission materials
scripts/ Release validation and packaging
docs/ Privacy, support, terms, publishing
See CONTRIBUTING.md and CODE_OF_CONDUCT.md. All changes must preserve dependency-free core tests and pass the release validator. The release process and Codex submission checklist are documented in docs/PUBLISHING.md.