PlugWiki

Create Technical Spike skill: time-boxed research documents for Claude Code

Makes your AI assistant write time-boxed spike documents that answer one technical question before you build.

Installs
9K
on skills.sh
Stars
40.7K
on GitHub
Updated
Oct 2026
last repo commit

create-technical-spike has your assistant write a spike document: one technical question, a time limit, a research plan and a recommendation. It saves each spike as a markdown file in docs/spikes (the default folder), named like api-copilot-integration-spike.md. The skill also tells the agent to search docs, read your codebase and build small prototypes before it writes the decision. It bundles no scripts, and it sits in GitHub's community awesome-copilot repo.

Good for

  • Developers who must settle a technical unknown before they start building
  • Tech leads who want every research task to end in a written decision
  • Teams that want a repeatable spike template with category, owner and timebox

Install create-technical-spike

Claude Code

  1. 1

    Download the skill folder from https://github.com/github/awesome-copilot/tree/HEAD/skills/create-technical-spike.

  2. 2

    Put it at ~/.claude/skills/create-technical-spike/ so that SKILL.md sits directly inside.

  3. 3

    Start a new session; the skill loads when its description matches the task.

Built from the package data we hold. Not yet tested by us.

What's in the skill

When it loads

Create time-boxed technical spike documents for researching and resolving critical development decisions before implementation.

Sections in SKILL.md

  1. Document Structure
  2. Summary
  3. Research Question(s)
  4. Investigation Plan
  5. Research Tasks
  6. Success Criteria
  7. Technical Context
  8. Research Findings
  9. Investigation Results
  10. Prototype/Testing Notes
  11. External Resources
  12. Decision
  13. Recommendation
  14. Rationale
  15. Implementation Notes
  16. Follow-up Actions

Key facts

Author
Awesome Copilot Community (author field of the technical-spike plugin)
Repo
github/awesome-copilot
License
MIT
Installs on skills.sh
8,997 as of 2026-11-02
GitHub stars
40,666 for the whole github/awesome-copilot repo, 2026-11-02
Last repo commit
2026-10-30
What is inside
1 file, SKILL.md; it bundles no scripts
Part of
The technical-spike plugin (this skill plus 1 agent) and the project-planning plugin (8 skills, 7 agents)
Needs
SKILL.md names no packages or keys to install; the spike folder defaults to docs/spikes

What it includes

  • Writes one spike file with a fixed front matter and 7 sections

    Creates a markdown file with front matter (title, category, status, priority, timebox, dates, owner, tags). The body has Summary, Research Question(s), Investigation Plan, Technical Context, Research Findings, Decision and Status History.

    Try: “Create a technical spike on whether we can stream audio through the VS Code extension host. Timebox one week.”

    Write · Spike document template

  • Names files [category]-[short-description]-spike.md

    Uses kebab-case names that show the category and the unknown, such as performance-audio-processing-latency-spike.md.

    Files · File naming

  • API Integration, Architecture, Performance, Platform, Security, UX

    Each spike is filed under one of 6 categories: API Integration, Architecture & Design, Performance & Scalability, Platform & Infrastructure, Security & Compliance, User Experience.

    Read · Six categories

  • Gather information, validate with prototypes, then decide

    Phase 1 searches docs and your codebase and reads external resources. Phase 2 builds focused prototypes and runs experiments. Phase 3 writes the recommendation and follow-up tasks.

    Runs code · Research strategy

  • One question per spike, timeboxed, evidence-based, with a clear outcome

    One question per spike, a time limit, concrete evidence before closing, a clear recommendation, dependency tracking, and an actionable outcome.

    Read · Six rules for agents

  • Names search, fetch/githubRepo, codebase, runTasks, editFiles, vscodeAPI

    The last section lists the tools the agent should use: searching, fetching external APIs and examples, reading the codebase, running tasks for prototypes, editing files and testing VS Code extension capabilities.

    Runs code · Tools usage list

From the official docs

Permissions

Medium risk

The skill writes files into your repo and tells the agent to build and run small prototypes. Review each prototype before you let the agent run it.

What it can do

  • Read your data
  • Create and change things
  • Delete things
  • Read-only mode

Watch out for

  • It builds and runs prototypes in your project

    The research strategy says to create focused prototypes and run targeted experiments. Check what the agent plans to run before you approve it. github.com

  • The Copilot CLI plugin adds an agent

    The technical-spike plugin also ships the research-technical-spike agent, whose tools list includes execute, edit, web and agent. The skill alone does not add it. github.com

  • It bundles no scripts

    The folder holds only SKILL.md, so the behaviour above comes from instructions the agent follows. github.com

Technical details
Turning on read-onlyNo read-only mode. The skill is built to create files.
Scopes

  • Writes spike markdown files under docs/spikes

  • The research strategy tells the agent to create prototypes and run experiments or tests

  • The Tools usage section names fetch/githubRepo for external APIs, libraries and examples

Questions

What is a technical spike?

A short, time-boxed piece of research that answers one technical question before you build. This skill has your assistant write that research up as a markdown document with a plan, findings and a decision.

How do I install create-technical-spike?

Use the install box at the top of this page. In Claude Code, copy the skill folder into ~/.claude/skills/ or .claude/skills/. In Claude on the web or desktop, zip the folder and upload it under Customize, then Skills. The box also has Codex, the skills CLI and a GitHub Copilot CLI plugin.

Does it need an API key?

The skill does not ask for one. It bundles no scripts and names no keys, though your assistant may fetch external pages while it researches.

Where does it save the spike documents?

In docs/spikes by default. The SKILL.md uses a folder input with that default, so you can ask for another folder.

Is it safe?

It writes files in your repo and, per its research strategy, creates prototypes and runs experiments. Read each command before you approve it, and keep the work on a branch. See the permissions section above.

Which plugin includes it?

In awesome-copilot, the technical-spike plugin holds this skill and the research-technical-spike agent. The project-planning plugin holds it with 7 other skills and 7 agents. Both are GitHub Copilot CLI plugins, so the install box lists the smaller one for Copilot CLI.

How is it different from a PRD or implementation plan skill?

A spike is a research document for one open question. The project-planning plugin that holds this skill also lists skills for PRDs and implementation plans, such as breakdown-feature-prd and create-implementation-plan.

Information

updated Nov 2, 2026

Repository
github/awesome-copilot
Folder
skills/create-technical-spike

github/awesome-copilotskills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 5
  1. create-technical-spike SKILL.md · github.com
  2. awesome-copilot README, Install a Plugin · github.com
  3. technical-spike plugin.json · github.com
  4. project-planning plugin.json · github.com
  5. research-technical-spike agent · github.com

Updated Nov 2, 2026.