AI Services Market › Use cases › Software & DevOps
Software & DevOpsDone with GitHub Copilot 2026-10-26
Run terminal commands via Copilot CLI subagents
A commenter uses Copilot CLI agents that execute terminal commands, including the /fleet command and subagent-style workflows.
- What They Did
- The commenter used GitHub Copilot CLI in several terminal sessions at once, usually on GPT-5.4 with xhigh reasoning and often in autopilot mode. Their workflow was to start in /plan, use commands like /fleet to spin up parallel subagents and /delegate for async cloud work, then run /review, manually inspect changes, and watch progress live in the linked VS Code chat pane.
- How It Went
- They said the setup let them send agents to tackle work they never had time to do before and iterate faster, but it still required review and feedback from a capable engineer. Another commenter noted parallel agents can collide on the same repo, with worktree suggested as the fix.
- Worth Knowing
- The clearest catch is supervision: even the original poster said it still needs review, and they were checking in every 10–15 minutes while multiple terminals ran.
Try it yourself with GitHub Copilot
Use the Copilot CLI to tackle [large task, e.g. update all API endpoints to use the new auth helper] in my [repo]. Split it into independent pieces, use /fleet or subagents to run them in parallel, and have each one run its own terminal commands and tests. Show me the split plan before starting, and when done, give me a combined summary of changes and test results, and don't commit or push anything without asking.
Do this with a skill.
These AI Services Market playbooks do this exact job.