Workflow Writing Standard

To write concise AI agent workflows, use templates instead of examples, eliminate prose, merge sequential steps, and adhere to specific file, table, and code block size limits.

Common anti-patterns in skill file creation include verbose preambles, repeated examples, and pre-filled data, which reduce token efficiency and complicate AI agent processing.

When workflow files exceed size limits, you should follow specific guidelines for handling content overages in file sizes, table rows, and code blocks to maintain token efficiency.

Your AI workflow likely requires optimization because it contains unnecessary verbosity, such as pre-filled example data and prose, rather than using token-efficient templates.

You should avoid pre-filled example data when creating skill files. Creating reusable templates ensures conciseness, improves token efficiency, and prevents over-building.

The best way to structure prompt engineering workflows is to merge sequential steps, avoid verbose preambles, and enforce strict size limits for files, tables, and code blocks.