gsd-next

To automatically advance a multi-project workflow, the system reads STATE.md, ROADMAP.md, and phase directories to determine and trigger the next logical step autonomously through a no-argument invocation.

State-reading workflow orchestration inspects project state indicators like STATE.md and ROADMAP.md to detect current progress, then uses Codex adapters to orchestrate and execute the next cross-agent workflow phase without manual prompting.

No, you do not need to pass arguments to trigger the next workflow phase; the system operates autonomously on a no-argument model, inspecting state files to decide the appropriate next action automatically.

Yes, you can use this workflow automation across multiple project directories by reading STATE.md, ROADMAP.md, and phase directories within the environment to determine the next action for each specific project context.

The best way to orchestrate cross-agent workflow steps without manual prompting is using a state-reading automation model that evaluates project files and triggers subsequent phases via Codex adapters autonomously.

Workflow automation may fail to determine the next step if required state indicators like STATE.md, ROADMAP.md, or phase directories are missing or improperly formatted, preventing the system from reading the current project state.