Using PushBackLog with GitHub Copilot (MCP)

PushBackLog ships a built-in MCP (Model Context Protocol) server that integrates with GitHub Copilot in VS Code. This lets you read and write your backlog, decisions, diary entries, and time logs directly from your IDE using natural language.

What is MCP?

MCP is an open protocol that lets AI assistants (like GitHub Copilot) call tools on external servers. The PBL MCP server exposes your backlog as a set of callable tools that Copilot can use while you’re coding.

Setup

1. Generate an API Key

In PushBackLog: Settings → Under the Hood → API Keys → New API Key. Name it something descriptive (e.g. “Copilot — dev machine”) and copy the key.

2. Configure VS Code

Add a .vscode/mcp.json file to your project:

{
  "servers": {
    "pushbacklog": {
      "type": "http",
      "url": "https://your-api-host/mcp",
      "headers": {
        "x-api-key": "<your-api-key>"
      }
    }
  }
}

Replace your-api-host with your PBL API URL and <your-api-key> with the key you generated.

3. Reload the MCP Server

In VS Code, open the Command Palette (Cmd/Ctrl+Shift+P) and run MCP: Reload Servers.

Available Tools

Once connected, Copilot can use these tools:

Tool	Description
`list_projects`	List all your projects
`list_backlog_items`	List items in a project
`get_backlog_item`	Get a single item by ID or ticket key
`create_backlog_item`	Create a new item
`update_backlog_item`	Update fields, status, or structured content
`add_diary_entry`	Add a note/observation to an item
`list_diary_entries`	Read all diary entries on an item
`add_time_entry`	Log time spent on an item
`list_time_entries`	See time log for an item
`record_decision`	Create a new decision
`list_decisions`	List all decisions
`list_themes`	List themes/epics
`assign_theme`	Associate an item with a theme
`list_personas`	List team personas
`create_idea`	Create a new idea in the Idea Log
`list_ideas`	List ideas, optionally filtered by status
`get_idea`	Fetch a single idea by ID
`update_idea`	Update idea status, title, description, or tags

Typical Workflow with Copilot

Ask Copilot to pick up a ready ticket: “List all ready tickets and pick the highest priority one”
Copilot calls list_backlog_items and update_backlog_item to mark it in-progress
Copilot implements the feature, runs type checks, and commits the code
Copilot calls add_time_entry to log the time spent
Copilot calls update_backlog_item to move the ticket to in-review
Copilot calls add_diary_entry to summarise what was done

Security

API keys are scoped to your tenant and cannot access other organisations’ data
All MCP calls are authenticated and logged in the AI audit trail
Revoke a key immediately from Settings → API Keys if it is compromised