skill-name

To create reusable TRAE skills, define a SKILL.md file containing valid YAML frontmatter and a descriptive body. This packages deterministic agent instructions into a unit that TRAE loads on demand across projects.

On-demand loading discovers and activates skills only when task descriptions match, saving tokens and context. This mechanism reduces redundancy by injecting deterministic guidance solely when the specific automation task is triggered.

No external scripts are mandatory to package a TRAE skill. A valid skill requires only a SKILL.md frontmatter specifying the name and description, alongside a descriptive body, though optional resources directories can be included.

You standardize automation workflows by defining deterministic agent instructions as a reusable TRAE skill. Packaging these instructions in a SKILL.md ensures consistent execution and reduces redundancy across multiple projects.

The frontmatter requirements for a valid TRAE skill mandate specifying the skill name and description in YAML format. This frontmatter combined with a descriptive body enables skill discovery and on-demand loading.

TRAE skills fail to load if the SKILL.md frontmatter lacks mandatory name and description fields, or if the task description does not match the skill's descriptive body, preventing on-demand discovery and context injection.