spike

Time-boxed spike tickets are created in GitHub by running the /spike command with a short title. This generates a standard issue template with mandatory fields to capture your hypothesis, budget, and kill criteria before work begins.

A spike ticket is a standardized, time-boxed GitHub issue used for product backlog exploration to prevent scope drift. You need one to capture research questions, technical feasibility, architecture experiments, and risk assessments before committing to a feature.

You enforce mandatory fields and branch naming for spike tickets using a helper workflow and SKILL.md frontmatter. This setup requires explicit budget, kill criteria, labels, and disposition to be filled before work begins.

Yes, the spike ticket workflow operates without external dependencies. It relies entirely on GitHub Issues, standard templates, labels, and a helper workflow to enforce mandatory fields and branch naming for your research tasks.

The best way to prevent scope drift on technical feasibility questions is using a structured spike ticket. It requires explicit budget, kill criteria, and disposition upfront, ensuring clear decision points and fast archiving of research results.

You should not use a time-boxed spike ticket for tasks lacking a clear hypothesis or budget. Spikes require explicit kill criteria and disposition upfront, making them unsuitable for open-ended research without defined exit criteria.