output-style-builder

Define a JSON schema specifying your agent's required fields, data types, and validation rules. This Skill guides you through schema creation, embedding it in your agent configuration, and validating outputs against it to guarantee 97–100% compliance with your API or downstream system requirements.

Use a scope hierarchy—global, project, skill, and command levels—to cascade style rules. This Skill helps you set up centralized style definitions in .claude/styles/ and override them selectively at lower scopes, ensuring tone and format consistency while minimizing redundant configuration.

Yes. This Skill supports JSON, YAML, XML, and Markdown-KV formats. Select the format that best fits your use case—JSON for APIs, YAML for config files, XML for document exchange, or Markdown-KV for human-readable structured text—and configure validation gates for your chosen format.

This Skill embeds validation scoring with an 80-point minimum threshold. It checks output structure against your schema, applies constrained decoding to enforce compliance, and flags failures before delivery, guaranteeing reliable machine-parseable data from your agent.

Use format configuration when you need machine-readable, validated outputs for API integration, data pipelines, or downstream automation. Documentation alone cannot enforce schema compliance; this Skill adds validation gates and constrained decoding to eliminate inconsistency and parsing errors.

Inline embedding places format definitions directly in agent instructions; references store them in .claude/styles/ for reuse across multiple agents. This Skill helps you choose based on token cost, maintainability, and scope—references reduce token overhead and simplify updates across projects.