Back to directory

OpenPencil Skill

An LLM skill that teaches an AI coding agent how to design with OpenPencil, covering its command-line tool, a batch design DSL, MCP design tools and general design conventions.

OpenPencil Skill

The skill follows the agentskills.io specification and installs into Claude Code, Cursor, Codex, OpenCode or Gemini CLI, after the separate op command-line tool has been installed and connected to a running OpenPencil desktop app or web server. Once installed, asking an agent to design something with OpenPencil causes it to generate PenNode design documents through the op CLI or MCP tools automatically.

Included material covers the full op CLI reference for app control, document operations, nodes, import, export, templates, layout, pages, variables and code generation, a batch design DSL with a sandboxed JavaScript mode, the PenNode schema for the current OpenPencil version, more than 40 semantic component roles with sensible defaults, general design principles for typography, color and spacing, layout and sizing rules, a layered MCP workflow that goes from a skeleton to content to refinement, and support for importing SVG, HTML pages and captured web snapshots.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.