spike

Executes timeboxed exploratory coding work with a running journal of attempts and outcomes.

Updated Apr 13, 2026
One-click install
npx skills add https://github.com/scoots31/engineering-playbook --skill spike-scoots31
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: spike
Source: https://github.com/scoots31/engineering-playbook/tree/main/skills/spike
Command: npx skills add https://github.com/scoots31/engineering-playbook --skill spike-scoots31

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Exploratory coding often sprawls into unscoped work with no record of what was tried or learned. This Skill keeps spikes focused on a single outcome within a timebox and captures a running journal so the learning survives even when the code does not. ## Core Features & Use Cases - Timeboxed Exploration: Enforces a defined timebox with at most one 50% extension before forcing a land decision. - Running Journal: Maintains a dated spike-journal.md recording what was tried, what broke, what surprised, and the current best guess. - Graduation Triggers: Detects signals like sharing requests, auth needs, or file sprawl that mean the work has outgrown spike territory. - Use Case: You want to test whether a new API can power your feature. Run a two-day spike, journal each attempt, then hand the code and journal to the land step for a Keep, Toss, or Promote decision. ## Quick Start Start a spike to test whether the payments API supports recurring charges, timeboxed to one day, and keep a journal of everything you try.

Frequently Asked Questions about spike

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a timeboxed coding spike?

Define the outcome, timebox, and graduation triggers in a scope-check first, then work in as few files as possible while logging every attempt in a spike-journal.md. When the timebox ends, land the spike with a Keep or Toss decision.

What is a spike in agile software development?

A spike is timeboxed exploratory work aimed at answering a question or proving feasibility, not building production code. It produces a working attempt plus a journal of what was tried, which is mined later for decisions.

Should I write tests during a spike?

No. Tests, abstractions, and documentation are deliberately excluded during a spike because they signal the work has left exploratory territory. Hardcode values, inline everything, and let the journal serve as the record.

When should a spike be promoted to a real project?

Promote when graduation triggers fire: someone wants to use or share the work, you need auth or a second file, or the timebox has been extended repeatedly. Stop spiking and run the land step with Promote intent.

What happens when a spike timebox expires?

If the outcome was reached, land with a Keep or Toss decision. If it was almost reached, one extension of up to 50 percent of the original timebox is allowed. If unreachable, land with Toss and capture the learning in the journal.