SkillShop
SkillShop › Use cases › Software & DevOps
Software & DevOpsDone with custom-built or unnamed agent Sep 15, 2026

Turn one goal into an ordered plan of coding-agent tasks

Ordewell takes a single goal and breaks it into an ordered plan of tasks for coding agents.

Turn one goal into an ordered plan of coding-agent tasks preview
What They Did
You give Ordewell a goal. A planner, which can be a coding agent running read-only, researches the repo and asks clarifying questions. It then commits a plan where each task has its own runner, model, effort and mode. You can edit the plan before running it. On run, it starts a real coding-agent session per AI task, in parallel where dependencies allow, and passes each task its predecessors' results.
How It Went
In the README's example, a rate-limiting goal became a four-task plan (three AI, one manual) that finished with 4 completed, 0 failed. A task passes only when its completion marker appears in the runner's output. No benchmarks are given.
Worth Knowing
The TUI needs tmux, and Windows users must run it under WSL. Parallelism defaults to 3 concurrent sessions (configurable from 1 to 5).
Try it yourself with custom-built or unnamed agent

I have one goal: [describe your goal]. Read my repo at [path] and break this goal into an ordered plan of small coding tasks, noting which tasks depend on others and what done looks like for each. Save the plan as PLAN.md and stop there without changing any code, so I can review it first.

Read the original ↗