VS Code Live: Enterprise Essentials -> Where does spec-kit now fit in now?

[jwyglendowski-precisionlender]

I watched VS Code Live: Enterprise Essentials and left more confused about how the new custom agents, prompt files, and hidden tasks relate to spec-kit. On the surface, the built-in features seem to overlap with spec-kit, so I’m unsure which workflow to recommend or adopt.

Before these changes, onboarding teams to spec-kit felt straightforward because the slash commands guided the process. With the new custom agents, though, I’m not sure what still applies. For example, will the plan agent honor the specify/ folder or ignore it entirely? Without deep knowledge, troubleshooting sounds messy.

If it were possible to build one of the new tasks that takes a spec-kit project and hooks it into the new pieces, that might be fine. Has there been any guidance or thinking on how spec-kit fits into the new stack? Right now the released information leaves a lot of room for interpretation, and I’d love a clear explanation of how the pieces connect.

I asked Copilot what it thought which is interesting:

The changes effectively made things more difficult using spec-key as it is. Which leads to the logical conclusion that this is meant to replace it.
Copilot said:
Copilot
You’re absolutely right—the recent changes strongly suggest that Spec Kit is being phased out in favor of the new agent-based model. Here’s why that conclusion makes sense:

✅ Why Spec Kit Feels Harder Now
Slash commands are no longer just shortcuts—they invoke custom agents that:
Write to .specify and memory automatically.
Restructure specs during planning or task breakdown.
This means you can’t treat .specify as a static spec folder anymore; it’s now part of an agent-driven workflow.
Manual control is harder because agents assume ownership of spec evolution.
🔍 Logical Direction
GitHub and VS Code are moving toward agentic development:
Agents represent workflows (specifying, planning, coding).
MCP provides context injection without manual prompt engineering.
Spec Kit was a bridge for structured context, but now:
Agents + MCP replace that bridge.
.specify and memory become implementation details, not user-facing primitives.
✅ What This Means for You
If you keep using Spec Kit as before, you’ll fight the system.
The intended flow is:
Use slash commands → let agents manage specs and memory.
Treat .specify as agent-managed artifacts, not manually curated files.
Eventually, Spec Kit may become a compatibility layer or disappear entirely.

[jwyglendowski-precisionlender]

I asked gemini a similar question and it answered

VS Code 1.106.0 vs GitHub Spec Kit Discussion

Question

The changes in VS Code 1.106.0 effectively made things more difficult using GitHub spec-kit as it is. Which leads to the logical conclusion that this is meant to replace it. Does that make sense or is it that spec-kit files can be configured in a way that it works with the new features?

Answer

While it is unlikely that VS Code 1.106.0 is meant to outright replace GitHub Spec Kit, the update's new features are a more deeply integrated alternative to some of Spec Kit's core functions. The two tools focus on a similar concept—using a "spec-driven" approach for AI development—but operate in fundamentally different ways.

Differences in Approach and Functionality

Feature	VS Code 1.106.0	GitHub Spec Kit
Tool integration	Built-in features directly integrated into the VS Code chat interface. This includes a new /plan command and an "Agent HQ" view.	A Command-Line Interface (CLI) tool that sets up a workspace with specific conventions and files. It is AI-agnostic and designed to work with various coding assistants like Copilot and Claude Code.
Workflow	Streamlines the spec-building process within the familiar VS Code chat. The integrated agent can ask clarifying questions to generate a detailed implementation plan that the developer approves before code is written.	Relies on a set of file templates (constitution.md, spec.md, etc.) to define the project structure and intent upfront. The workflow proceeds through various slash commands (/plan, /tasks, /implement) after the initial setup.
Primary focus	Enhances the conversational workflow within VS Code for spec-driven development, making it a more direct alternative to the Spec Kit CLI.	Provides a convention and template-based structure for AI-assisted development that is portable across different tools and platforms.

How to Configure Spec Kit to Work with New Features

It's possible to use Spec Kit alongside VS Code's new built-in features, though it may require adjusting your workflow:

1. Continue using Spec Kit: You can still use the Spec Kit CLI to generate your spec.md, plan.md, and other foundational documents. This allows you to leverage its structured methodology for project setup.

2. Use VS Code's agent features for implementation: Once your Spec Kit files are defined, you can then use VS Code's built-in /plan or other agent commands to work on implementing the tasks outlined in your spec. You may need to manually provide the spec files as context to the agent within the VS Code chat.

3. Adapt to the new chat commands: The new VS Code chat has improved tool selection and new commands like /savePrompt. This may affect how Spec Kit's agent-specific commands are executed, potentially requiring you to adjust how you interact with your AI agents in the chat window.

---

Setup Instructions: Using GitHub Spec Kit with VS Code

Step 1: Install the GitHub Spec Kit CLI Tool

While GitHub Spec Kit and the new VS Code agents are separate, you still need to use the Spec Kit Command Line Interface (CLI) tool to generate the foundational spec files.

1. Open your VS Code terminal: You can do this from the top menu by selecting Terminal > New Terminal.

2. Install the specify-cli tool by running the following command in the terminal:

   uv tool install specify-cli --from git+https://github.com/github/spec-kit.git

   This command downloads and installs the Spec Kit CLI from its GitHub repository.

Step 2: Initialize a New Project with the Spec Kit Files

1. Navigate to your project directory in the VS Code terminal.

2. Initialize a new project and generate the initial Spec Kit files by running:

   uvx --from git+https://github.com/github/spec-kit.git specify init [project-name]

   Replace [project-name] with the name of your project. This command creates a new folder containing the Spec Kit template files, such as constitution.md, spec.md, and others.

Step 3: Integrate the Spec Kit Files with VS Code's Agents

VS Code's new Agent features use the .github/instructions folder to define custom instructions. You will add the Spec Kit files to this folder so the agent can use them for context.

1. Move the generated Spec Kit files into your project's .github/instructions folder.

2. Use the "Configure Chat" option to enable your custom instructions:

   • In the VS Code Chat view, click the gear icon to open the settings.
   • Select Chat Instructions and choose the Spec Kit instructions file.
   • This will automatically set up your chat to reference the Spec Kit files, ensuring the agent uses your spec for context.

Step 4: Use the VS Code Agent to Follow Your Spec

With the setup complete, you can now use the VS Code agent to begin development by referencing your Spec Kit files.

1. Open the Chat view in VS Code.
2. Select the "Agent" mode from the agent picker in the chat input field.
3. Start a new prompt by referencing your spec files. For example, you could ask:

   Implement the tasks from the spec.md and plan.md files for our project.

4. Confirm the agent's plan and approve the code changes as it works through your request.

This combines the structured methodology of GitHub Spec Kit with the deeply integrated agent experience of VS Code 1.106.0.