Trainer Train - Skill

Use this skill as the orchestration contract for one trainer run against an agent skill target: a SKILL.md file and its associated directory of references, scripts, assets, and evals.

Read references/skill-loop-contract.md for the full routing table, spec-compliance rules, frontmatter optimization constraints, and body-content validation requirements before any stage execution.

When to use this skill

• The selected target is a SKILL.md file or a full skill directory.
• The caller needs to initialize or resume a trainer workspace for a skill target.
• The improvement focus is triggering accuracy (frontmatter), execution reliability (body), structure, or all three.
• Missing datasets or eval manifests need to be synthesized before optimization.
• A winning candidate needs to be validated against spec compliance and written back.

Do not use this skill for raw prompt files, Python code targets, or agent contracts. Read the parent trainer skill's references/target-routing.md to identify the appropriate specialist for those target types.

Required inputs