| name | artbot-cli |
| description | Use when operating the ArtBot CLI for local backend lifecycle, health checks, run creation, run inspection, replay debugging, review queue work, canaries, cleanup, or command-first automation. Prefer JSON-first commands, explicit subcommands, and non-TUI usage. Relevant terms include ArtBot CLI, artbot doctor, backend status, research artist, research work, runs show, runs watch, replay attempt, review queue, canaries, cleanup, and command-only automation. |
ArtBot CLI
Use this skill when you need to operate ArtBot effectively from the command line.
Core Rules
- In this repo, prefer
pnpm --filter artbot dev -- ....
- For installed usage, replace that prefix with
artbot ....
- Bare
artbot opens the interactive UI in an interactive TTY. Use explicit subcommands plus --json for automation.
- For agent work, prefer explicit commands and
--json or --output-format stream-json.
- Keep
--no-tui or ARTBOT_NO_TUI=1 available when automation must refuse interactive UI launch.
- Interactive setup, TUI launch, auth capture, and local backend start/stop require
artbot trust allow for the current workspace.
Decision Tree
If you are not sure whether the environment is healthy:
pnpm --filter artbot dev -- --json doctor
pnpm --filter artbot dev -- --json backend status
pnpm --filter artbot dev -- --json auth list
pnpm --filter artbot dev -- --json auth status
pnpm --filter artbot dev -- trust status
If the machine is not bootstrapped yet or local config is broken:
pnpm --filter artbot dev -- setup
If you want to understand planned sourcing before creating a run:
pnpm --filter artbot dev -- --json research artist --artist "Burhan Dogancay" --preview-only
pnpm --filter artbot dev -- --json research work --artist "Erol Akyavas" --title "Kusatma" --preview-only
If you want the finished terminal payload in one command:
pnpm --filter artbot dev -- --json research artist --artist "Burhan Dogancay" --wait
pnpm --filter artbot dev -- --json research work --artist "Erol Akyavas" --title "Kusatma" --wait
If you already have a run id:
pnpm --filter artbot dev -- --json runs show --run-id <id>
pnpm --filter artbot dev -- --json runs watch --run-id <id> --interval 2
pnpm --filter artbot dev -- --output-format stream-json runs watch --run-id <id>
pnpm --filter artbot dev -- sessions resume
If parsing or acceptance looks wrong, replay stored artifacts before re-running:
pnpm --filter artbot dev -- --json replay attempt --run-id <id>
If a run produced review items or clustering questions:
pnpm --filter artbot dev -- --json review queue --run-id <id>
pnpm --filter artbot dev -- --json review decide --run-id <id> --item-id <item-id> --decision merge
pnpm --filter artbot dev -- --json graph explain --run-id <id> --cluster-id <cluster-id>
If you need smoke checks or retention work:
pnpm --filter artbot dev -- --json canaries run
pnpm --filter artbot dev -- --json canaries history
pnpm --filter artbot dev -- --json cleanup --dry-run
Usage Notes
--json is the default choice when another tool or agent will read the output.
--output-format stream-json is preferred when another tool wants incremental NDJSON lifecycle events.
setup is intentionally interactive; do not use it as a normal health probe.
artbot tui is for humans. Stay in command mode unless the task explicitly calls for the interactive UI.
- Respect auth and licensed access boundaries. Do not attempt bypass behavior.
- Install this repo skill explicitly by copy or symlink from
skills/artbot-cli; package install must not write into agent homes automatically.