---

name: project-cli-and-sessions description: "Create and manage Kedro projects through the CLI, project metadata, settings, sessions, packaging, notebooks, and safe diagnostics." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

Project CLI and Sessions

Use this sub-skill when a task is about creating a Kedro project, identifying whether a directory is a Kedro project, using global or project-specific kedro commands, configuring project metadata/settings, bootstrapping a project for Python code, using KedroSession, packaging a project, or opening project-aware IPython/Jupyter sessions.

Route Here

• Create projects with kedro new, --config, --tools, --example, --starter, --checkout, --directory, or --telemetry.
• Diagnose CLI command availability, project detection, pyproject.toml metadata, project layout, starter/template behavior, and telemetry prompts.
• Use bootstrap_project(), configure_project(), KedroSession.create(), session.load_context(), project settings.py, kedro package, kedro ipython, or kedro jupyter.
• Prepare a packaged-project command such as python -m <package_name> or programmatic package-mode startup.

Route Elsewhere

• Pipeline graph authoring, node(), Pipeline, pipeline(), namespacing, tags, or kedro pipeline create/delete design choices: read pipelines and nodes.
• kedro run execution semantics, runner selection, node slicing, resume, and KedroSession.run() filtering behavior: read runners and execution.
• Catalog, credentials, parameters, config loaders, and data/config validation: read data catalog and config.
• Hooks, plugins, custom CLI commands, custom datasets, and extension entry points: read hooks and extensions.
• Inspection snapshots or optional HTTP server behavior: read inspection and server.
• Package-wide install/import or router guidance: read the Kedro root skill.

Current Facts

• Kedro version target: 1.4.0; distribution and import name: kedro; Python requirement: >=3.10.
• Console script: kedro = kedro.framework.cli:main; python -m kedro is also supported.
• Global commands include info, new, and starter; project commands include catalog, ipython, jupyter, package, pipeline, registry, run, and server.
• KedroSession.create(project_path=None, save_on_close=True, env=None, runtime_params=None, conf_source=None) creates a one-successful-run session.
• bootstrap_project(project_path) reads project metadata, adds the source directory to import paths, and calls configure_project(package_name).
• configure_project(package_name, preserve_logging=False) configures settings, pipeline registry, package name, and project logging for package mode.

Reference Map

• Read CLI reference for exact global/project commands, kedro new flags, project command routing, telemetry-safe command forms, and optional notebook dependencies.
• Read project workflows for project creation, starter/template choices, metadata/layout requirements, settings defaults, packaging, and notebook/IPython workflows.
• Read sessions and context for bootstrap_project(), configure_project(), KedroSession, context loading, runtime parameters, config source overrides, and package-mode startup.
• Read troubleshooting when CLI commands are missing, project detection fails, starter creation errors, settings/session/bootstrap fails, optional dependencies are absent, or telemetry prompts block automation.
• Run kedro_cli_probe.py to perform safe kedro --version, kedro --help, and kedro info checks with telemetry opt-out and a read-only current-directory Kedro project probe.

Fast Patterns

• Prefer telemetry-safe probes in automated workflows: KEDRO_DISABLE_TELEMETRY=1 kedro --version, KEDRO_DISABLE_TELEMETRY=1 kedro --help, and KEDRO_DISABLE_TELEMETRY=1 kedro info.
• Create a non-interactive project with explicit choices: KEDRO_DISABLE_TELEMETRY=1 kedro new --name="My Project" --tools=lint,test --example=n --telemetry=no.
• Use starters only when Git/network access is acceptable, or use a local starter path; official aliases and remote VCS starters may invoke Cookiecutter/Git.
• Check project detection before project commands: Kedro looks for pyproject.toml containing [tool.kedro] in the current directory or a parent directory.
• For Python scripts outside a project command, call bootstrap_project(project_root) before KedroSession.create(project_path=project_root).
• For installed package mode, call configure_project(package_name, preserve_logging=True) when repeated configuration must not remove runtime logging handlers.

Safety Notes

• kedro new writes a new project directory; run it from the intended parent directory or pass an explicit output_dir through a config file.
• --starter, --tools=pyspark, or --example=yes may fetch templates through Git/Cookiecutter; warn users before networked or remote-template workflows.
• kedro package builds artifacts under dist/ and requires packaging tools such as build; it also archives configuration while excluding local/*.yml.
• kedro ipython and kedro jupyter require optional interactive dependencies; missing modules are an install issue, not a project-detection issue.