CLI Reference

ignition <command>

Commands

Apply a recipe to target hosts. Use --check for a dry-run.

ignition run <recipe> <targets> [options]

Argument	Description
`recipe`	Path to recipe `.ts` file
`targets`	Comma-separated targets: host names, `@group`, `user@host:port` (e.g. `@web,@db`)

Load and print an inventory file.

ignition inventory [file] [options]

Argument	Description
`file`	Path to inventory `.ts` file

Options:

Flag	Default	Description
`--inventory <file>`	—	Path to inventory file (alternative)
`--trace`	`false`	Accepted by the CLI, but currently has no effect

Scaffold a new Ignition project in the current directory.

ignition init

Creates recipe.ts, inventory.ts, and ignition.config.ts, ensures package.json exists, and adds @grovemotorco/ignition if needed.

Start a persistent web dashboard server.

ignition dashboard [options]

Options:

Flag	Default	Description
`--host <host>`	`127.0.0.1`	Dashboard hostname
`--port <port>`	`9090`	Dashboard port
`--max-history <n>`	`10`	Number of historical runs to retain

Display schema and resource information. Useful for tooling and integrations.

ignition schema <command>

Subcommands:

These also appear in current CLI help because they come from the underlying CLI framework:

Command	Description
`completions`	Generate shell completion script
`mcp add`	Register Ignition as an MCP server
`skills add`	Sync generated skill files

Flag	Description
`--check`	Dry-run without applying changes

Flag	Default	Description
`--inventory <file>`	—	Path to inventory file
`--host-key-policy <policy>`	`accept-new`	`strict`, `accept-new`, or `off`
`--identity <path>`	—	Path to SSH private key
`--no-multiplex`	—	Disable SSH connection multiplexing

Flag	Default	Description
`--error-mode <mode>`	`fail-fast`	`fail-fast`, `fail-at-end`, or `ignore`
`--tags <tag>`	—	Filter recipes by `meta.tags` (repeatable)
`--var <key=value>`	—	Set a variable (repeatable)
`--confirm`	—	Prompt before applying
`--parallelism <n>`	`5`	Max concurrent hosts
`--host-timeout <ms>`	`0`	Per-host timeout (0 = unlimited)
`--resource-timeout <ms>`	`30000`	Per-resource timeout
`--retries <n>`	`2`	Retry attempts for transient failures
`--retry-delay <ms>`	`1000`	Initial retry backoff

Flag	Default	Description
`--trace`	—	Show SSH commands and detailed output

--tags currently matches the recipe module's export const meta = { tags: [...] }. Repeat the flag instead of comma-joining values:

ignition run deploy.ts @web --tags web --tags deploy

Flag	Default	Description
`--cache`	—	Cache check results across runs
`--cache-ttl <ms>`	`600000`	Cache entry lifetime
`--cache-clear`	—	Clear cache before running

Flag	Default	Description
`--dashboard-host <host>`	`127.0.0.1`	Dashboard server hostname to report to
`--dashboard-port <port>`	`9090`	Dashboard server port to report to

Flag	Default	Description
`--log-dir <path>`	—	Directory for NDJSON audit logs

Flag	Description
`--filter-output <keys>`	Filter output by key paths
`--format <toon\|json\|yaml\|md\|jsonl>`	Output format
`--help`	Show help
`--llms`	Print the LLM-readable manifest
`--mcp`	Start as an MCP stdio server
`--schema`	Show JSON Schema for a command
`--token-count`	Print token count instead of output
`--token-limit <n>`	Limit output to `n` tokens
`--token-offset <n>`	Skip the first `n` tokens
`--verbose`	Show the full output envelope
`--version`	Print version and exit

Code	Meaning
`0`	Success — command completed successfully
`1`	Failure — resource failures or other CLI/config/runtime errors

Variable	Description
`DENO_DEPLOY_TOKEN`	Token for sandbox microVMs
`SSH_AUTH_SOCK`	SSH agent socket (used by system SSH)