PROJECT MANAGEMENT

AI Scope-Creep Triage: Classify Late Additions and Route or Park

When an issue is added to an in-progress Linear sprint, an LLM judges whether it is a true emergency or deferrable scope creep.

CategoryProject Management
EngineSim + Paperclip
Difficultyadvanced
Triggerevent
Steps5
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerIssue added to active cycle (Linear webhook)LinearLinear
  • ActionClassify emergency vs. scope creep with LLMOpenAI
  • LogicBranch on classification verdict
  • ActionEscalate emergency to PM in SlackSlack
  • OutputPark deferrable issue in backlog with commentLinearLinear

What it does

This workflow intercepts every issue dropped into a running Linear cycle and uses an LLM to classify it: genuine same-sprint emergency (production blocker, customer-down) versus deferrable work that is really scope creep. Emergencies get escalated; deferrable items are automatically parked in the backlog with an explanatory comment.

When to use it

Use it on high-traffic teams where a human PM cannot review every mid-sprint addition fast enough. It applies a consistent triage rule so urgent work flows through while nice-to-haves stop silently bloating the committed scope.

How it works

  1. 1A Linear webhook fires when an issue enters the active cycle after its start.
  2. 2An OpenAI step reads the issue title, description, and labels and returns a classification with a confidence score and reasoning.
  3. 3A logic branch splits on the verdict.
  4. 4If 'emergency', an action posts an escalation to the PM Slack channel with the model's rationale for sign-off.
  5. 5If 'deferrable', an action removes it from the cycle, moves it to the backlog, and adds a Linear comment explaining the auto-park and how to override.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect LinearIssues, projects, cycles, triage.
  2. 2
    Connect OpenAIModels, embeddings, files.
  3. 3
    Connect SlackChannels, DMs, threads, mentions.
  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.