Technical experiment planner

What it does

Bring one technical unknown and a time limit. Get a study plan with proof needed to choose a path.

What goes in and out

You give it

• The blocking decision, affected code, constraints and deadline.
• A named owner and realistic research or prototype time budget.

You get

• A focused spike document with tasks, dependencies and completion criteria.
• A findings/decision record that separates planned experiments from actual results.

How the task works

• Fix the question and timebox. Choose one technical decision, its downstream impact, owner and deadline.
• Design decisive evidence. Specify document/code research and the smallest prototypes or tests that discriminate options.
• Close with a bounded decision. Record observed results, recommendation, limits and concrete follow-up work; leave unperformed tests planned.

Full instructions

• Open your AI chat, or your AI coding assistant if you want it to search code and run tests.
• Paste the skill text, then add your blocking question, the code it touches, your time limit, the owner and the deadline.
• Ask your AI to write the spike document. Check that it has one question only.
• Run the research tasks yourself, or let your AI do them if it can run code. Fill in the findings.
• Write the decision, then turn the follow-up actions into implementation tasks.

Full instruction file

Create Technical Spike Document

Create time-boxed technical spike documents for researching critical questions that must be answered before development can proceed. Each spike focuses on a specific technical decision with clear deliverables and timelines.

Before you start

Use the material below to do this job. If facts or tool access are missing, explain what you can finish and what must wait.

What you need (use your own material or the linked skill):

• Your question (yours): the decision or problem, and what a good outcome looks like.
• A codebase (yours): a project's code, in a folder or repo the AI can open.

How to start:

1. Look first. Use what they've already attached, pasted or put in their about-me note. In a coding assistant with file access, look in the project too for these by name.
2. Check what's missing. Look for your question, a codebase. They can attach files, paste the details, or share a file path your assistant can read. Ask only for missing facts that change the result. If they ask you to start with what they have, complete the parts you can and mark the gaps.
3. Do the job. Follow the skill below with the facts and tools you have. Mark each gap with MISSING: where it belongs and never invent a fact to fill it.

Document Structure

Create individual files in ${input:FolderPath|docs/spikes} directory. Name each file using the pattern: [category]-[short-description]-spike.md (e.g., api-copilot-integration-spike.md, performance-realtime-audio-spike.md).

---
title: "${input:SpikeTitle}"
category: "${input:Category|Technical}"
status: "🔴 Not Started"
priority: "${input:Priority|High}"
timebox: "${input:Timebox|1 week}"
created: [YYYY-MM-DD]
updated: [YYYY-MM-DD]
owner: "${input:Owner}"
tags: ["technical-spike", "${input:Category|technical}", "research"]
---

# ${input:SpikeTitle}

## Summary

**Spike Objective:** [Clear, specific question or decision that needs resolution]

**Why This Matters:** [Impact on development/architecture decisions]

**Timebox:** [How much time allocated to this spike]

**Decision Deadline:** [When this must be resolved to avoid blocking development]

## Research Question(s)

**Primary Question:** [Main technical question that needs answering]

**Secondary Questions:**

- [Related question 1]
- [Related question 2]
- [Related question 3]

## Investigation Plan

### Research Tasks

- [ ] [Specific research task 1]
- [ ] [Specific research task 2]
- [ ] [Specific research task 3]
- [ ] [Create proof of concept/prototype]
- [ ] [Document findings and recommendations]

### Success Criteria

**This spike is complete when:**

- [ ] [Specific criteria 1]
- [ ] [Specific criteria 2]
- [ ] [Clear recommendation documented]
- [ ] [Proof of concept completed (if applicable)]

## Technical Context

**Related Components:** [List system components affected by this decision]

**Dependencies:** [What other spikes or decisions depend on resolving this]

**Constraints:** [Known limitations or requirements that affect the solution]

## Research Findings

### Investigation Results

[Document research findings, test results, and evidence gathered]

### Prototype/Testing Notes

[Results from any prototypes, spikes, or technical experiments]

### External Resources

- [Link to relevant documentation]
- [Link to API references]
- [Link to community discussions]
- [Link to examples/tutorials]

## Decision

### Recommendation

[Clear recommendation based on research findings]

### Rationale

[Why this approach was chosen over alternatives]

### Implementation Notes

[Key considerations for implementation]

### Follow-up Actions

- [ ] [Action item 1]
- [ ] [Action item 2]
- [ ] [Update architecture documents]
- [ ] [Create implementation tasks]

## Status History

| Date   | Status         | Notes                      |
| ------ | -------------- | -------------------------- |
| [Date] | 🔴 Not Started | Spike created and scoped   |
| [Date] | 🟡 In Progress | Research commenced         |
| [Date] | 🟢 Complete    | [Resolution summary]       |

---

_Last updated: [Date] by [Name]_

Categories for Technical Spikes

API Integration

• Third-party API capabilities and limitations
• Integration patterns and authentication
• Rate limits and performance characteristics

Architecture & Design

• System architecture decisions
• Design pattern applicability
• Component interaction models

Performance & Scalability

• Performance requirements and constraints
• Scalability bottlenecks and solutions
• Resource utilization patterns

Platform & Infrastructure

• Platform capabilities and limitations
• Infrastructure requirements
• Deployment and hosting considerations

Security & Compliance

• Security requirements and implementations
• Compliance constraints
• Authentication and authorization approaches

User Experience

• User interaction patterns
• Accessibility requirements
• Interface design decisions

File Naming Conventions

Use descriptive, kebab-case names that indicate the category and specific unknown:

API/Integration Examples:

• api-copilot-chat-integration-spike.md
• api-azure-speech-realtime-spike.md
• api-vscode-extension-capabilities-spike.md

Performance Examples:

• performance-audio-processing-latency-spike.md
• performance-extension-host-limitations-spike.md
• performance-webrtc-reliability-spike.md

Architecture Examples:

• architecture-voice-pipeline-design-spike.md
• architecture-state-management-spike.md
• architecture-error-handling-strategy-spike.md

Best Practices for AI Agents

1. One Question Per Spike: Each document focuses on a single technical decision or research question

2. Time-Boxed Research: Define specific time limits and deliverables for each spike

3. Evidence-Based Decisions: Require concrete evidence (tests, prototypes, documentation) before marking as complete

4. Clear Recommendations: Document specific recommendations and rationale for implementation

5. Dependency Tracking: Identify how spikes relate to each other and impact project decisions

6. Outcome-Focused: Every spike must result in an actionable decision or recommendation

Research Strategy

Phase 1: Information Gathering

1. Search existing documentation using search/fetch tools
2. Analyze codebase for existing patterns and constraints
3. Research external resources (APIs, libraries, examples)

Phase 2: Validation & Testing

1. Create focused prototypes to test specific hypotheses
2. Run targeted experiments to validate assumptions
3. Document test results with supporting evidence

Phase 3: Decision & Documentation

1. Synthesize findings into clear recommendations
2. Document implementation guidance for development team
3. Create follow-up tasks for implementation

Tools Usage

• search/searchResults: Research existing solutions and documentation
• fetch/githubRepo: Analyze external APIs, libraries, and examples
• codebase: Understand existing system constraints and patterns
• runTasks: Execute prototypes and validation tests
• editFiles: Update research progress and findings
• vscodeAPI: Test VS Code extension capabilities and limitations

Focus on time-boxed research that resolves critical technical decisions and unblocks development progress.

After this

When you've handed it back, suggest what they could do next in one short question. It's a suggestion: if they're done, that's fine. Offer these, and say they can type their own:

• The Research project runner. It takes the document. If they pick one and that skill is added here, carry straight on with what you just made. If not, give the link and say what to bring.