opencode-agent-generator

This Skill generates OpenCode agent files with complete YAML frontmatter, including description, model, temperature, tools, and permission controls. It automates agent creation in project folders with proper syntax validation and kebab-case naming, eliminating manual configuration errors.

You can configure tool permissions (read, write, bash) based on agent purpose. The Skill applies role-specific permissions—a security auditor might have audit tools, a code analyst gets read and grep access, enabling secure and functional agents without overprivileging.

The Skill integrates swarm communication protocols directly into generated agents, defining interaction guidelines and responsibilities. This ensures seamless collaboration between multiple agents without manual protocol definition or coordination setup.

Yes. The Skill creates domain-specific agents for roles including code analysis, documentation writing, security auditing, and test engineering. Each agent receives a focused system prompt and tool permissions tailored to its domain responsibilities.

Agents are created project-locally in the `.opencode/agent/` directory with kebab-case naming and YAML frontmatter. This placement keeps agents version-controlled alongside your project and ready for use with `opencode serve`.

No. The Skill automatically generates focused system prompts based on the agent's purpose and role. You specify the domain and responsibilities, and it produces a production-ready prompt with appropriate tone and task boundaries.