OpenWork Enterprise

OpenWork Enterprise is the OpenWork Factory superproject: one workspace for product repos, operational automation, and reusable OpenCode behaviors.

The aim is simple: make teams dramatically faster by turning repeated work into predictable, auditable workflows.

Why this exists

• Build automation-first workflows, skills, and runbooks that scale beyond ad-hoc scripts.
• Coordinate development and operations across multiple related repos without losing context.
• Capture company operations as code + docs so processes are reproducible, inspectable, and improvable.

What lives here

• AGENTS.md - behavior contract for the OpenWork Factory assistant in this workspace.
• ISSUES.md - recurring conceptual and ops issues to keep fixing systematically.
• .opencode/skills/ - reusable capabilities (syncing submodules, worktree flow, testing, release helpers, etc.).
• .opencode/commands/ - command shortcuts such as the release flow.
• _repos/ - pinned product and support repos, including openwork, opencode, opencode-browser, opencode-scheduler, agent-mint, agent-bank, and agent-watch.
• _worktrees/ - isolated worktrees for task-by-task implementation.
• prds/ and research/ - planning and research artifacts.

Default working loop

1. Sync root + submodules before starting work.
2. Use a dedicated worktree for each task.
3. Implement in the correct repo under _repos/.
4. For user-facing or remote behavior, run the OpenWork Docker stack and verify through Chrome MCP.
5. Keep docs/prompts/skills updated as behavior changes.

Quick start

1. Clone this repo.

2. Initialize submodules:

   git submodule update --init --recursive

3. Read AGENTS.md (workspace behavior) and _repos/openwork/AGENTS.md (product source of truth).

4. Use the existing skills and commands in .opencode/ to run repeatable workflows.

Current focus areas

• GitHub issues triage
• PR update triage
• Service domain setup and package maintenance
• Notion CRM and documentation
• Better release evidence and post-release verification