Decompose proposals into feasibility questions and run disposable implementation spikes.
npx skills add https://github.com/JamesFincher/gengar --skill spike-jamesfincher
Please help me install this Agent Skill. Skill: spike Source: https://github.com/JamesFincher/gengar/tree/main/skills/software-development/spike Command: npx skills add https://github.com/JamesFincher/gengar --skill spike-jamesfincher
This skill solves the risk of wasting time building the wrong solution by validating feasibility through small, disposable experiments before committing to a real implementation.
Use case: You’re considering two competing ways to parse PDFs for extraction; run two parallel spikes (e.g., pdfjs vs camelot), compare results, and decide which approach to invest in.
Tell the agent: "Spike out whether my proposed approach is feasible, including 2–5 concrete feasibility questions and an observable prototype, then give me a VALIDATED/PARTIAL/INVALIDATED verdict."
High-intent search queries and answers about installing and using this skill.
To validate a software idea before building, decompose the proposal into 2–5 concrete feasibility questions framed with Given/When/Then, run disposable implementation spikes, and produce evidence-based verdicts to de-risk your MVP plan.
A spike in software engineering is a time-boxed, disposable experiment used to probe technical unknowns and de-risk MVP plans. Use a spike when you need comparative testing of approaches or observable evidence of feasibility before committing to full production work.
To compare approaches for feasibility, run parallel disposable spikes for each option, build a minimal runnable artifact demonstrating behavior, and evaluate the evidence to decide which approach to invest in for the real build.
Yes, you can test MVP feasibility without a complete application by doing minimal targeted research per spike and building a minimal interactive artifact that demonstrates behavior, yielding a VALIDATED, PARTIAL, or INVALIDATED verdict.
The best way to de-risk an MVP plan with technical unknowns is to break the proposal into observable feasibility questions, present an ordered spike table, build disposable prototypes, and generate evidence-based verdicts before full production work begins.
Feasibility questions for rapid prototype testing should be structured using Given/When/Then syntax so each outcome is observable, allowing the spike to demonstrate clear behavior and produce an evidence-based recommendation for the real build.
carterwayneskhizeineDecompose ambiguous ideas into risk-ordered Given/When/Then feasibility spikes with verdicts.
GitnappValidate uncertain technical ideas through focused throwaway experiments before production implementation.
anxiety135790Decompose ideas into testable spike questions and validate feasibility.
tae0yVerify behavior through public APIs in unit and integration tests.
