Add another feature from the backlog (agentic loop)

Goal#

Pick up an additional feature from the Octocat Supply backlog and deliver it with a fuller agentic loop — plan → implement → test → review. With the cart underway, the team needs to keep burning down the backlog before launch.

Actions#

• Choose a backlog item — Pick a meaningful feature from assets/backlog.md (e.g. payment integration for the cart, order history, product search/filtering, or inventory management). A set of these may already be seeded as GitHub Issues in your repo — check the Issues tab and pick one.
• Plan with an agent — Use GitHub Copilot to break the work into steps and identify the files and layers involved (frontend, API, database) before writing code.
• Implement — Build the feature across the stack, keeping changes consistent with existing patterns and type-safe.
• Add tests — Cover the new logic with unit tests (API/repository) and component tests where relevant.
• Review — Use Copilot to review your own change before considering it done.

Success criteria#

• A new, working backlog feature is delivered end to end and demonstrable in the running app.
• The feature has appropriate automated test coverage.
• Code follows existing repository conventions and is type-safe.
• You used an agentic plan → implement → test → review loop and can point to the agent's plan, the implementation, the tests, and the review pass.

Optional stretch — harnesses, MCP & requirement refinement#

• Compare harnesses — Run the planning step twice: once in Copilot CLI and once in the Copilot App / IDE chat. Note which gave the more implementation-ready plan for a quick fix vs. a multi-file change, and write down a decision rule.
• Wire in an MCP server — Connect one MCP server (e.g. the GitHub MCP server) to pull real backlog context such as issues and PRs. Note what stays local-only vs. what the server can see.
• Refine a requirement from organizational signal — Point your agent at assets/workiq/ — the mock WorkIQ organizational knowledge (Teams thread, support-ticket digest, stakeholder email). Have it synthesize the signal across those artifacts, separate a genuinely-new need from noise and from work already in the backlog, and shape one unmet need into an issue-ready requirement: problem statement → acceptance criteria → scoped tasks (optionally opening it as a GitHub Issue). There's no answer key — the skill is the refinement. A ready-made Requirement Refiner custom agent (.github/agents/) can guide this.

Learning resources#

• About GitHub Copilot coding agent
• Extending Copilot Chat with the Model Context Protocol (MCP)
• GitHub MCP server