See llms.txt for all machine-readable content.

Back to Templates

Rank and triage dependency risk with GitHub, OSV, Slack, npm, PyPI and OpenAI

Created by

Created by: Daniel Shashko || tomax
Daniel Shashko

Last update

Last update 14 hours ago

Categories

Share

Quick overview

This workflow runs weekly to scan GitHub repositories for JavaScript and Python dependencies, checks their release health via npm/PyPI and known vulnerabilities via OSV, then ranks the highest-risk packages and posts a single Slack digest and (when needed) opens a GitHub issue for replacements.

How it works

  1. Runs every Monday morning on a schedule.
  2. Reads package.json and a configured Python manifest (for example pyproject.toml or requirements.txt) from each listed GitHub repository and extracts dependency names and versions, normalizing version ranges to bare numbers.
  3. Looks up each dependency in npm or PyPI to capture the latest version, last release date, deprecation/yank status, and basic maintenance signals.
  4. Queries OSV for advisories affecting the specific dependency version and filters out findings that are already fixed, flagging advisories with no fix available.
  5. Calculates a risk score and verdict (replace/upgrade/watch/fine) from staleness, deprecation, severity, and fix availability, then builds a ranked shortlist.
  6. Uses an OpenAI-compatible chat model to generate one actionable sentence per shortlisted package.
  7. Upserts the results into an n8n Data Table, posts a single summary message to Slack, and creates one GitHub issue when any packages are marked for replacement.

Setup

  1. Add a GitHub credential with access to the repositories you want to scan and the repository where issues should be created.
  2. Add a Slack credential and set the target channel for the digest message.
  3. Add an OpenAI-compatible model credential/configuration for the chat model used to write one-line recommendations.
  4. Create an n8n Data Table named dependency_risk with columns package_key, verdict, risk, summary, and last_reported.
  5. Update the Settings values for repos (comma-separated owner/name), python_manifest, issue_owner, issue_repo, and any thresholds like max_packages, shortlist, and staleness day limits.

Requirements

  • A GitHub credential that can read the repositories you list. npm, PyPI and OSV take no credential at all.

Customization

  • Point python_manifest at whichever file your repos actually use - pyproject.toml, requirements.txt, or a path of your own. Both PEP 621 and Poetry layouts are parsed.