Skip to content
Back to skills

Runner And Programmatic

ASecurity

"Guides programmatic Metaflow execution with Runner, NBRunner,

  • 247 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 8, 2026
devopspythongoapi

Works with

  • cli
  • api

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned September 8, 2026

npx -y skills add VectorSpaceLab/AREX-Skill --skill runner-and-programmatic --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Runner And Programmatic?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Runner And Programmatic
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/vectorspacelab-runner-and-programmatic/badge)](https://www.skillsdirectory.com/skills/vectorspacelab-runner-and-programmatic)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: runner-and-programmatic
description: "Guides programmatic Metaflow execution with Runner, NBRunner,
  Deployer bridge objects, subprocess status, logs, async waits, and returned
  run/task inspection."
disable-model-invocation: true
metadata:
  disco-role: operating
license: Apache 2.0
---

# Runner and Programmatic Execution

Use this sub-skill when a task asks to run a Metaflow flow from Python code, automate a local smoke test, inspect `ExecutingRun`/`ExecutingTask`, stream logs, or bridge to deployment objects programmatically.

## Quick Route

- Read [`references/runner-api.md`](references/runner-api.md) for verified `Runner` usage, kwargs placement, returned objects, status, logs, and artifact access.
- Read [`references/notebook-and-spin.md`](references/notebook-and-spin.md) for notebook and Spin-adjacent APIs.
- Read [`references/troubleshooting.md`](references/troubleshooting.md) when subprocesses fail, time out, or return an unexpected `Run` lookup error.
- Run [`scripts/runner_smoke.py`](scripts/runner_smoke.py) for a self-contained local programmatic smoke test.

## Minimal Programmatic Pattern

```python
from metaflow import Runner

with Runner("flow.py", show_output=False, pylint=False, env={"USERNAME": "disco"}) as runner:
    executing = runner.run(max_workers=1)
    print(executing.status, executing.returncode)
    print(executing.run.pathspec)
```

Top-level flow-script options such as `pylint=False`, `profile=...`, `env=...`, or `datastore=...` belong on `Runner(...)`. Run-command options such as `max_workers=1`, `tags=[...]`, or `run_id_file=...` belong on `runner.run(...)`.

## Boundaries

- For writing the flow file itself, route to [`../flow-authoring/SKILL.md`](../flow-authoring/SKILL.md).
- For detailed object traversal after a run exists, route to [`../client-and-data/SKILL.md`](../client-and-data/SKILL.md).
- For real Argo or Step Functions deployment prerequisites, route to [`../deployment-orchestration/SKILL.md`](../deployment-orchestration/SKILL.md).

Files in this skill

  • SKILL.md2 KB
  • references/notebook-and-spin.md1.1 KB
  • references/runner-api.md1.6 KB
  • references/troubleshooting.md1.1 KB
  • scripts/runner_smoke.py1.9 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.