GitHub’s legal team built custom workflow tools with Copilot CLI — no manual coding. The GitHub blog post walks through how a non-engineering team used the CLI to turn scattered notes and one-off prompts into reusable, structured workflows.
What Copilot CLI gives non-dev teams:
- Custom agents — convert ad-hoc prompts into repeatable, shareable workflows that standardize how a task gets done regardless of who runs it
- Autonomous execution — the CLI applies changes without the GUI overhead of clicking “Keep” or “Undo” on every projected edit
- Git-diff review — changes land in the repo and get reviewed through diffs, the way developers already review code
- Parallel tasks — fleet mode runs multiple operations concurrently
- Image input — drag-and-drop or paste visual references for UI-adjacent work
The legal team angle is the real story. Lawyers think in contracts and compliance, not Python. Copilot CLI’s custom agents let them encode institutional knowledge — how to handle a data request, what a compliance check requires — into structured, repeatable processes instead of relying on whoever happens to be on shift.
NOTE
Custom agents are built for exactly this pattern: taking knowledge that lives in scattered notes and one-use prompts and turning it into shared workflows. Teams where the same task gets approached differently depending on the operator get consistency for free.
The barrier isn’t technical skill — it’s knowing what workflow you want to automate. The CLI handles the code; the legal team handles the domain knowledge.