Prioritise project issues with AI
Turn a crowded backlog into decisions about what starts now, what needs evidence and what can be closed with a reason.
Prioritise project issues with AI when every open report has a label but the team still does not know what to tackle first. The issue-triage skill looks at impact, urgency and likely effort, groups duplicate reports and distinguishes a missing fact from a low priority. The output includes the work the team deliberately declines.
Imagine a broken order form, an intermittent export problem and a cosmetic request. Count who is affected, whether a workaround exists and what evidence is available. The intermittent issue may need a sample file before its priority can be judged. Assigning that request for evidence is more useful than a guessed ranking.
What it does
- Combine duplicates while retaining how often a problem was reported.
- Separate an operational blocker from an inconvenient but usable path.
- Limit current work to what the team can actually start.
- Give incomplete reports an owner and a concrete information request.
- Record why a request is deferred, redirected or closed.
Install and use
- Use an agent with an authorised issue list, impact evidence and an effort estimate where available. Tracker access is needed only to apply decisions there.
- Ask for a ranked set of actions, including a reason for anything that should be closed.
- Review assignments and closures with the accountable owner before updating a shared tracker.
FAQ
Source
Made by Arjun Prabhulal. Original skill: issue-triage, published on GitHub under the MIT licence.
View on GitHub