tech-spike

Runs time-boxed technical spikes to validate assumptions and produce a structured spike report.

Updated Apr 17, 2026
One-click install
npx skills add https://github.com/mistakenot/skills --skill tech-spike-mistakenot
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: tech-spike
Source: https://github.com/mistakenot/skills/tree/main/plugins/exploration/skills/tech-spike
Command: npx skills add https://github.com/mistakenot/skills --skill tech-spike-mistakenot

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Committing to an implementation based on untested assumptions leads to costly rework. This Skill de-risks ideas by running focused, time-boxed experiments that produce concrete evidence before you build. ## Core Features & Use Cases - Assumption Identification: Generates and ranks 3-7 high-risk assumptions (integration, performance, compatibility, correctness, feasibility) for user confirmation. - Isolated Spike Execution: Runs all exploratory code inside a gitignored .tmp/ directory, using parallel sub-agents to validate independent assumptions concurrently. - Evidence-Based Reporting: Writes a SPIKE-REPORT.md with a go/no-go verdict, per-assumption results with confidence levels, surprises, risks, and reproduction steps. - Use Case: Before building a feature on a new library, spike whether the library supports your use case, whether performance holds at expected scale, and whether edge cases are handled — then decide with evidence. ## Quick Start Ask the agent to run a tech spike on your idea, for example: "Spike whether we can stream large files through this API without hitting memory limits."

Frequently Asked Questions about tech-spike

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

FAQPage Schema
How do I run a tech spike before starting implementation?▼

Describe the idea or assumptions you want tested, and the skill generates a ranked list of high-risk assumptions, designs validation approaches for each, executes them in an isolated .tmp/ workspace, and writes a SPIKE-REPORT.md with a go/no-go verdict.

What is a tech spike in software development?▼

A tech spike is a time-boxed exploration to answer a specific technical question or validate an assumption. This skill formalizes the process: define assumptions, design validations, execute throwaway experiments, and document evidence-based findings.

Does spike code pollute my main project repository?▼

No. All spike work happens inside a .tmp/ directory at the project root, which the skill automatically creates and adds to .gitignore if missing. Spike code is throwaway and never touches the main project tree.

Can multiple assumptions be tested in parallel?▼

Yes. Independent assumptions are validated concurrently by separate sub-agents, each with its own working directory under .tmp/spike-NNN-name/. Each sub-agent captures what it tried, observed results, and whether the assumption held.

When should I not use a tech spike?▼

Skip spikes when requirements are well understood and the approach is proven. Spikes answer specific high-risk questions; they are not a substitute for building production code, writing tests, or doing full design work.