onboard

Generate role-based project onboarding documentation by scanning CLAUDE.md, role definitions in .opencode/agents/, and relevant directories to produce a tailored markdown file summarizing architecture, conventions, and first tasks for a specific contributor role.

Onboard programmers or narrative writers by scanning targeted repository areas like src/ or design/narrative/ to extract engineering standards, current sprint context, dependencies, and common pitfalls into a structured role-specific onboarding document.

Yes, you can create an onboarding document for QA and production roles by scanning relevant directories such as tests/ and production/ to summarize project state, role-specific priorities, and an actionable first tasks plan.

Role-specific contributor onboarding works by reading role definitions from .opencode/agents/ and correlating them with scanned repository areas to output a markdown file detailing what a new team member should read and do first.

Yes, a CLAUDE.md file is required because the onboarding process reads it to summarize the current project context, architecture, and conventions before generating the role-based onboarding markdown file.

Limitations include dependency on existing CLAUDE.md and .opencode/agents/ files for context, and the onboarding document generation requires user confirmation before writing the final markdown file to the repository.