spike

What problem does it solve? Before investing in a real implementation, teams often need to know whether an idea is technically feasible, which of several approaches works best, or where hidden risks lie. This Skill structures that exploration as disposable spike experiments so unknowns are answered with working code instead of guesswork. ## Core Features & Use Cases - Decompose ideas into feasibility questions: Break a vague idea into 2-5 independent spikes framed as Given/When/Then statements, ordered by risk so the idea-killing question runs first. - Comparison spikes: Build competing approaches side by side (e.g., pdfjs vs camelot for PDF parsing), optionally in parallel via delegated subagents, then produce a head-to-head verdict table. - Structured verdicts: Every spike directory ends with a VALIDATED, PARTIAL, or INVALIDATED verdict documenting what worked, what didn't, surprises, and recommendations for the real build. - Use Case: You want to know if LLM tokens can stream to a browser client over WebSockets in under 100ms. The Skill creates spikes/001-websocket-streaming/ with a runnable prototype, tests edge cases, and records a verdict before you commit to the architecture. ## Quick Start Ask the agent to spike whether your idea is feasible, for example: "Spike whether we can stream LLM tokens over WebSockets with under 100ms latency before I commit to this architecture."