generate command

For the step-by-step workflow, see Generate Tasks from the CLI.

generate

Generate tasks from user stories using a template. By default runs as a dry run (preview only). Use --execute to actually create tasks.

Usage

atomize generate [template] [options]
atomize gen [template] [options]  # alias

Arguments

Argument	Description
[template]	Path to a YAML template file or catalog ref (e.g. template:backend-api). If omitted, you will be prompted.

Options

Option	Type	Default	Description
-p, --platform <platform>	string	azure-devops	Platform to use. Options: azure-devops, mock
--profile <name>	string	-	Named connection profile to use (see auth add)
--execute	flag	-	Actually create tasks (default is dry-run preview)
--auto-approve	flag	-	Required with --execute in non-interactive mode to acknowledge live task creation
--continue-on-error	flag	-	Keep processing other stories if one fails
--story <ids...>	string[]	-	Target specific work items by ID, bypassing the template filter. excludeIfHasTasks still applies.
--story-concurrency <n>	number	3	Max stories processed in parallel (max: 10)
--task-concurrency <n>	number	5	Max tasks created in parallel per story (max: 20)
--dependency-concurrency <n>	number	5	Max dependency links created in parallel (max: 10)
--limit <n>	number	-	Cap the number of work items processed (useful for testing before a full --execute run)
-v, --verbose	flag	-	Show detailed output including per-task breakdown
-o, --output <file>	string	-	Write a JSON report to this file path (for CI/CD)
--include-sensitive-report-data	flag	-	Include descriptions, custom fields, and platform-specific work item data in the JSON report (--output only)
-q, --quiet	flag	-	Suppress non-essential output
--json	flag	-	Print the JSON report to stdout instead of the human-readable output. Runs a dry run unless --execute is also passed. Implies --quiet
--json-stream	flag	-	Print newline-delimited JSON progress events, then a final report event, to stdout. Implies --quiet; cannot be combined with --json

The concurrency options are hidden from atomize generate --help.

Examples

Dry run (default: no --execute):

atomize generate template:backend-api

Execute for real:

atomize generate template:backend-api --execute

Target specific stories by ID (bypasses filter):

atomize generate template:backend-api --story STORY-123 STORY-456

Use --story to skip the template's filter criteria and process explicit work items directly. Useful for:

• Re-running a generation that failed partway through
• Processing a one-off story that does not match the template's usual filter
• Testing the template against a known story before a full run

excludeIfHasTasks still applies: stories that already have tasks are skipped unless you remove that flag from the template.

Execute for real in CI/non-interactive mode:

atomize generate template:backend-api --execute --auto-approve

Mock platform (no credentials needed):

atomize generate template:backend-api --platform mock

Test against a subset of items before a full run:

atomize generate template:backend-api --limit 5

CI/CD with JSON report:

atomize generate template:backend-api \
  --execute \
  --output report.json \
  --quiet

By default the JSON report omits story descriptions, custom field values, and raw platform payloads to avoid accidentally writing sensitive data to disk or log collectors. Pass --include-sensitive-report-data to include them:

atomize generate template:backend-api \
  --execute \
  --output report.json \
  --include-sensitive-report-data

When to use --include-sensitive-report-data:

• Debugging unexpected task output (custom field values, description interpolation)
• Auditing exactly what was sent to the platform
• Internal tooling that processes the full report programmatically

Do not use it in shared CI pipelines where the report artifact may be visible to other users.

Exit Codes

Code	Meaning
0	All stories processed successfully, including a dry run
1	One or more stories failed, or live execution was refused
2	No stories matched the filter or --story IDs
3	Authentication failed

The exit codes are the same with --json and --json-stream. A run with no matching stories exits 2 and still prints its report, with storiesProcessed: 0.