Operational features that keep the workstream moving.

Scripts are reusable shell commands you can register inside the app and run against a selected workstream repo or a custom path. Output is streamed into an integrated terminal, and executions can be cancelled, inspected, and reused later.

Polystro includes template-driven workflows for actions like review, generation, fixes, refactors, tests, and documentation. These runs let you choose a workstream, branch, commit, CLI, and model before execution.

The key distinction is that template-driven runs can stage proposed actions and approvals, which makes them more structured than raw chat prompts and more flexible than simple shell scripts.

The same settings area also distinguishes Polystro-managed /commands from raw CLI passthrough //commands, which is how reusable prompt files and direct CLI command handling can coexist in one shell.

Scheduled automations build on the same template-driven idea, but they run on a schedule. The app tracks whether an automation is enabled, when it ran last, and when it will run next.

Polystro exposes operational configuration directly in Settings. That includes skill registration and sync, MCP server configuration, custom slash commands, notifications, keybindings, voice settings, editor preferences, CLI detection, API key overrides, model routing, and routing policies.

Skills and MCP are especially important if your agent workflows depend on shared local instructions or external tool servers across supported CLIs.

The local viewer is Polystro's way to continue a workstream away from the main desktop surface. The app can open a loopback viewer locally or expose a LAN share link over HTTPS when available.

Devices can pair by QR code or a Mobile/Web link and PIN on first enrollment, then reconnect later as trusted devices. Trusted device access is visible in Settings, where each device can be revoked individually.