update-agent

To create and update AI agent templates automatically, use a tool that scans your designated template directory, decides whether to create or modify files, and writes the system prompt body with YAML frontmatter enforcing name and description fields.

YAML frontmatter is metadata at the top of a Markdown file that enforces required fields like name and description for agent templates. It defines the agent's identity and capabilities before the system prompt body is written.

To add a new specialized agent, create a Markdown file in the agent templates directory with a descriptive name, include YAML frontmatter with name and description, then provide the agent system prompt in the body.

Yes, you can delete or deprecate old agents from your template directory. The update process handles deletion scenarios for deprecated agents, modifying the template directory to reflect current requirements and removing obsolete files.

Yes, updating an agent template modifies the system prompt body. The template body is written as the system prompt, and changes to requirements trigger modifications to both the YAML frontmatter and the body content.

The best way to manage agent lifecycle changes across a template directory is automating the process of locating existing agents, deciding whether to create, update, or delete them, and persisting changes to the correct template path.