Use when installing or upgrading the `helix` binary, when a `helix` command fails to authenticate or hangs on an organization prompt, when no first-class subcommand exists for an endpoint, or when managing Helix organizations, teams, members, secrets or…
Use when moving files into or out of Helix — uploading to the filestore, building or re-indexing a knowledge/RAG base, attaching logs or screenshots to a spec task, or copying files in and out of a running sandbox container.
Use when running, watching or steering a Helix coding agent — starting a spec task, chatting with a running task, reading its conversation, exec'ing or copying files into its container, screenshotting or streaming its desktop, debugging a stuck agent, or…
Build, configure, troubleshoot and evaluate helix-org bots (org agents) — especially customer-facing support bots that run as limited-permission bot instances behind an app-key gateway. Covers writing the bot's prompt and runbook, shipping tested page/form…
Use when defining, deploying, chatting with or testing a Helix agent (app) — applying agent YAML, wiring API/OpenAPI tools, MCP servers or knowledge into one, granting team or user access, running YAML-defined evaluations, or registering models and provider…
Use when an agent is asked for an interactive page, dashboard, visualization, prototype, report, PDF, image, shareable HTML, or compiled static SPA that should be uploaded to and served by a Helix project rather than left in a sandbox, or when updating,…
Use when working with a Helix project or its Kanban board — creating or forking a project, applying project YAML, listing or filtering the board, creating cards, moving them between columns, setting priority, labels or assignees, approving specs, archiving…
Use when installing, self-hosting or upgrading Helix, deploying it to Kubernetes, or setting up a sandbox/Hydra runner node; and when a deployment misbehaves — spec tasks stuck without a sandbox, a node online but never placed on, desktops black or dying,…