Top 200 ReposEXPLORER
Browse agents, categories and pages
Back to results
Reported by sourceThe source says this happened. Top 200 Repos has not reproduced it.

Trial spec-driven development with spec-kit on an existing app

Birgitta Böckeler tried GitHub's spec-kit workflow with Copilot in VS Code on a mid-sized dashboard feature and found the process heavy for the size of the task.

The problem

How it works

As part of comparing Kiro, spec-kit and Tessl, she ran spec-kit's slash-command workflow (starting with /specify) through GitHub Copilot in VS Code. The target was a new modal summarizing data from an existing dashboard, which she sized as a 3-5 point story with many dependencies on existing code.

Results and limitations

She reports the workflow generated many repetitive markdown files that were tedious to review, that the agent sometimes treated notes about existing classes as new specifications and duplicated them, and that plain AI-assisted coding would have been as fast. She did not finish the implementation.

Before you try it

One person, one unfinished feature, and tool versions from October 2025 that may have changed. It is an expert's qualitative assessment, not a measured study.

Tools and repositories

Tools
spec-kit · GitHub Copilot · VS Code

Try it yourself: Pilot spec-driven development on one small feature

Fill the fields to adapt the prompt. Empty fields stay as visible [placeholders].

Preview
We want to test whether spec-driven development fits our team. Using [Spec-driven tool] in [Repository path], take the feature described in [Feature brief] and draft only the specification and plan stages. Before writing anything, list the existing modules the feature touches and mark which parts of the spec describe existing behavior and which describe new behavior, so nothing gets regenerated. Keep the spec under [Maximum spec length (pages)] pages and flag any step that feels redundant. Do not write implementation code or modify existing files yet. At the end, estimate how long a reviewer needs for these documents compared with implementing [Feature brief] directly through normal agent-assisted coding.

Time-box the pilot and compare review effort honestly before adopting the workflow across a team.

Related videos

About github/spec-kit

About github/spec-kit

Independent creators. Listing a video implies no endorsement, sponsorship or partnership.

Sources

  1. Understanding Spec-Driven-Development: Kiro, spec-kit, and Tessl (opens in a new tab)

    First-person account · martinfowler.com · Published 15 Oct 2025 · Source checked 9 Nov 2026, 22:15 UTC

    What this source supports
    • spec-kit used with GitHub Copilot in VS Code
    • Feature was a dashboard summary modal sized at 3-5 points
    • Many repetitive markdown files, tedious to review
    • Agent duplicated existing classes as new specs
    • Implementation not finished
Report a problem

Reported by source

Turn a feature idea into a spec, plan and tasks before coding

Use Spec Kit's slash commands to have a coding agent write a spec, a technical plan and a task list, then implement against them.

github.comgithub/spec-kit
Reported by source

Apply the context engineering lessons from Manus to a long-running agent loop

The Manus team describes the context rules it settled on after rebuilding its agent framework four times: keep the prompt prefix cacheable, mask tools instead of removing them, and use files as memory.

manus.im · 18 Jul 2025
Reported by source

Build an app with Replit Agent while keeping production data out of reach

Replit Agent builds and browser-tests apps from a description. A widely reported 2025 incident, where it deleted a live database during a code freeze, shows why separate dev and prod data matters.

docs.replit.com
Reported by source

Explore and change a large codebase from the terminal with Gemini CLI

Gemini CLI is Google’s open-source terminal agent. Simon Willison found it strong on large codebases in 2025; in 2026 Google moved individual users to Antigravity CLI.

github.comgoogle-gemini/gemini-cli
Reported by source

Hand coding tasks to Cursor cloud agents that open pull requests

Start Cursor agents from the editor, web, Slack or a GitHub comment; each works in its own cloud VM on a branch and returns a pull request with screenshots, videos and logs.

cursor.com