The flow

[shukiv]

Hello,

Nice tool and I am using it a lot lately.
I would like to ask what is the correct flow using these skills to solve issues and designing an interface?

[wictorstenseke]

@shukiv
I use:

1. grill-me, 2. write-a-prd, 3. prd-to-plan and then executing with superpowers execute-plan skill. Trying out Matts tdd skill soon as well. (Not sure if skills names are 100% accurate)

[TDWeiTao]

Hi, guys 👋
Have you ever encountered a situation where code was generated using TDD, but the unit tests were not actually executed?

[mattpocock]

This was generated by AI during triage.

Thanks for asking — this is a real gap. Rather than answer it inline, the plan is to address it with documentation: a guide on how the skills compose into end-to-end flows (e.g. grill-me → PRD → to-issues → tdd). Keeping this issue open as a tracker for that docs work.

[mattpocock]

This was generated by AI during triage.

Reclassifying as ready-for-human.

Agent Brief (human-only)

Why this can't be delegated

This is a documentation-design task, not a mechanical doc edit. It requires the maintainer's judgment on which flows are the recommended ones (vs. merely possible), and how opinionated to be — an agent would either generate something generic or invent a flow the maintainer doesn't actually endorse.

What to write

A guide (likely a new top-level section in README.md, or a docs/flows.md) showing how the skills compose end-to-end. At minimum:

• A "happy path" flow from idea → shipped code (e.g. grill-me → to-prd → to-issues → tdd).
• 1–2 alternative flows for different starting points (e.g. starting from a bug report; starting from a vague idea).
• For each step: which skill, what it produces, what feeds the next step.

Acceptance

• A new reader can pick a starting point and follow a recommended flow without guessing which skill to use next.
• Linked from the top-level README.md.
• Doesn't enumerate every skill — just the ones on the recommended flows.

Open questions for the human

• One canonical flow, or multiple parallel ones?
• Include the third-party skills users mentioned in the thread (superpowers execute-plan, etc.) or stay strictly within this repo?