Frame and answer feasibility questions with throwaway spike code and a formal verdict.
npx skills add https://github.com/chughtapan/safer-by-default --skill spike-chughtapan
Please help me install this Agent Skill. Skill: spike Source: https://github.com/chughtapan/safer-by-default/tree/main/skills/spike Command: npx skills add https://github.com/chughtapan/safer-by-default --skill spike-chughtapan
Frame and answer a precise feasibility question with throwaway spike code and a formal verdict. The output is a written verdict with evidence, not production code. Use when the question is open-ended or when a quick go/no-go is required rather than a full implementation.
Frame a precise yes/no feasibility question, create a spike on a dedicated branch, and publish a structured verdict with evidence.
High-intent search queries and answers about installing and using this skill.
A spike is throwaway code used to test a single feasibility question and produce a formal go/no-go verdict. You need a spike when evaluating whether a library or approach is tractable within fixed time and budget, rather than building a full implementation.
To run a feasibility spike, frame one precise question, create a dedicated unmerged branch for throwaway code, execute the test, and publish a structured verdict with supporting evidence and a graduation plan if the approach proves viable.
No, you should never merge spike code into your main branch. Spikes are throwaway experiments designed to answer a feasibility question and produce a documented verdict, not to ship as production code.
After a successful feasibility spike, you follow a graduation path to architect or spec. You use the documented verdict and evidence to plan the real implementation, leaving the throwaway spike branch unmerged.
No, a proper spike follows a one-question discipline. You frame a single precise feasibility question to get a clear, evidence-driven go/no-go verdict, rather than testing multiple open-ended questions simultaneously.
A feasibility spike is ideal for evaluating a new library when you need a quick go/no-go decision. It provides an evidence-driven verdict with execution traces, avoiding the overhead of a full implementation when routing to deeper investigation is unnecessary.
me2reshCreates a time-boxed GitHub issue with kill criteria and promote/discard disposition.
wraps-teamRun a three-agent technical spike producing a feasibility verdict and execution plan.
Owl-ListenerDefine version naming, branching, and changelog workflows for design files.
XyHalcyonDecompose software ideas into feasibility questions and build disposable experiments.
