skill-creator

To scaffold a new Codex skill, run the initialization script to generate a standardized SKILL.md frontmatter, an OpenAI YAML agent file, and optional resource directories for scripts and references. This ensures consistent structure and reduces setup time.

The best way to bootstrap scalable Codex skills is using an automated scaffolding template that produces validated YAML agent configuration and metadata frontmatter. This provides a ready-to-use skeleton that teams can iterate on for new or updated skills.

Yes, the Codex skill initializer requires the pyyaml dependency to parse and generate the openai.yaml agent file required for UI integration and discovery. This ensures the YAML configuration is properly structured and validated.

You can update existing Codex skills as well as create new ones. The scaffolding process applies to both scenarios, covering metadata updates, resource directory additions for scripts and references, and agent configuration validation.

When initializing a skill for deterministic tasks, the scaffolding optionally creates resource folders for scripts, references, and assets. These directories provide structured locations for executable code, documentation, and supporting files.

Your new Codex skill needs frontmatter with name and description fields to ensure proper discovery and UI integration. The standardized SKILL.md template automatically generates these required metadata fields during the scaffolding process.