OpenASE is a ticket-driven automated software engineering platform whose core philosophy is "create a ticket, and leave the rest to the agent." It treats every piece of work as a ticket, and each ticket is bound to a workflow. AI agents automatically claim and execute tickets based on status triggers, ultimately delivering results with full traceability.
Purpose and Positioning OpenASE is aimed at software engineering scenarios that need to incorporate AI coding agents into a controllable process. It does not replace the agents themselves, but instead provides them with ticket orchestration, workflow execution, machine management, and observability, keeping humans in the control loop. It offers two complementary modes of human-agent interaction: when requirements are clear and a Harness exists (a hard-boundary document that constrains agent behavior), the ticket agent can autonomously complete the entire task; when requirements are ambiguous or exploration is needed first, synchronous dialogue can be conducted through Project AI.
Main Capabilities
- Ticket-driven orchestration: kanban and list views, parent-child tickets and dependency tracking, custom statuses and priorities, repository scope binding.
- Multi-agent support: Claude Code, Codex, Gemini CLI; real-time streaming output (SSE); agent lifecycle management; concurrency execution control.
- Workflow engine: Markdown Harness documents, skill binding and lifecycle hooks, scheduled tasks (Cron), built-in role templates.
- Machine management: local, direct connection, and reverse connection modes; WebSocket execution and SSH bootstrapping; health probing and resource metrics; connectivity diagnostics.
- Authentication and security: OIDC browser login (Auth0, Entra ID), agent platform token authentication, organization- and project-level RBAC, GitHub credential management.
- Observability: real-time activity event stream, agent run step tracking, GitHub Webhook integration, project update threads.
How to Use OpenASE is released as a single Go binary, with a built-in API server, workflow orchestrator, and embedded Web UI. It does not require Node.js at runtime (the SvelteKit frontend is compiled into the binary via go:embed). Installation methods include: a one-command installer (for brand-new Linux or macOS machines, automatically detecting the system, downloading and verifying the corresponding GitHub release package, providing PostgreSQL bootstrapping, and writing a runnable ~/.openase/config.yaml and ~/.openase/.env); it can also be built from source (perform the frontend build on a machine with sufficient memory; the control plane runs by default on http://127.0.0.1:19836）。配置工单状态、注册机器与代理、创建工作流后，即可观察代理执行。
Applicable Scenarios and Limitations Applicable scenarios include: converting clear requirements into autonomously executed ticket processes (full-stack coding agent or hybrid relay mode), conducting exploratory dialogue before submitting tickets, extending agent capabilities through skills, and multi-organization and multi-project management. Limitations and caveats: to maximize unattended execution, the platform starts CLI agents with permissive flags by default (Claude Code uses --dangerously-skip-permissions, Codex uses --yolo). Agents can read, write, and execute arbitrary commands on the host without per-operation confirmation; it is recommended to run only on machines whose agent access scope is trusted; the project is not designed for public internet deployment, and its target environments are local development, private networks, and trusted environments; browser access is always gated by authorization. Claude Code and Codex are recommended for production, while Gemini CLI is supported but less stable. Cross-organization team collaboration features are still under development.
