ENGINEERING

AI-triaged license violations with Slack approval escalation

When a PR introduces a flagged license, an agent assesses the actual compliance risk in context (license, package usage.

CategoryEngineering
EngineSim + Paperclip
Difficultyadvanced
Triggerevent
Steps6
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerGitHub PR opened or updatedGitHubGitHub
  • ActionCollect new dependencies, licenses, and usage contextGitHubGitHub
  • ActionAgent assesses compliance risk per packageOpenAI
  • LogicRoute allow, block, or needs-review
  • ActionEscalate gray-area cases to Slack with approve/rejectSlack
  • OutputUpdate PR commit status from the decisionGitHubGitHub

What it does

This workflow adds judgment to a license gate. When a pull request adds a package with a non-allowlisted or ambiguous license, an agent evaluates the real risk — considering the license family, whether the dependency is a dev-only tool, and the project's distribution model — then decides whether to hard-block immediately or escalate for human sign-off. Escalations go to a Slack channel with approve and reject actions, and the PR status reflects the outcome.

When to use it

Use this when a binary allowlist is too blunt — for example, weak-copyleft or dual-licensed packages that are sometimes acceptable. It routes the gray-area cases to legal or eng leadership instead of blocking everything outright.

How it works

  1. 1A GitHub pull_request webhook fires.
  2. 2The workflow gathers new dependencies, their licenses, and how each is used.
  3. 3The agent reasons about compliance risk and classifies each as allow, block, or needs-review.
  4. 4A logic branch routes block decisions to a failing PR status, and needs-review decisions to Slack.
  5. 5Reviewers approve or reject in Slack; the response updates the PR commit status accordingly.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect GitHubRepos, issues, pull requests, actions.
  2. 2
    Connect SlackChannels, DMs, threads, mentions.
  3. 3
    Connect OpenAIModels, embeddings, files.
  4. 4
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  5. 5
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  6. 6
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.