In an account published by XDA Developers on September 20, 2026, Nolen Jonker describes pairing two Claude Code skills: one to clarify a task before implementation, and another to guide work on the final deliverable. The author says this workflow helped sessions stay closer to the intended result—but it is a personal account, not evidence that skills guarantee completion.
Start by clarifying the task, not by building
Jonker’s first choice is the brainstorming skill in Jesse Vincent’s Superpowers project. Its role is to slow down the start of a task long enough to surface missing decisions. The described exchange is conversational: Claude asks one focused question at a time and waits for an answer before moving on.
![Claude AI for Beginners Bible: [5 in 1] The Ultimate Guide to Automate Your Work, Save Hours Every Week, and Use AI for Real-World Results](https://m.media-amazon.com/images/I/415+fSJacsL._SL160_.jpg)
As an Amazon Associate I earn from qualifying purchases.
Once the requirements are clearer, the skill helps develop a design for approval. This puts an explicit decision point before implementation: the user can correct assumptions or choose between approaches while the work is still being planned. Jonker says this made it less necessary to repeatedly redirect sessions toward the intended outcome; that is his reported experience, not a measured result.
Recommended Free Tools
After approval, turn the design into a plan
The next step is the separate writing-plans skill. In Jonker’s workflow, brainstorming does not write or execute the implementation plan: it first helps establish and approve a design. Writing-plans then turns that agreed design into a plan file with verification steps, giving the coding work a sequence and a way to check progress.
#1 Best Overall
- Make the request. State the task and relevant context.
- Answer focused questions. Resolve uncertainties before implementation begins.
- Review the design. Approve it or correct it while changes are still inexpensive.
- Use writing-plans. Create a plan from the approved design, including verification steps.
This sequence is useful when a task has unresolved requirements or several implementation choices. It is not necessary for every small change; for a clear, bounded request, the planning overhead may not be worthwhile.
Use an output skill for the deliverable
The second role is an output-oriented skill: guidance tailored to producing or editing a particular kind of artifact. Jonker points to Anthropic’s document skills, especially DOCX and PowerPoint examples. These are distinct from the opening brainstorming skill: they address how to handle the deliverable, rather than what the task should do.
For a document, a format-aware procedure can focus edits on the requested material. For slides, the cited example includes reviewing layouts, not merely generating slide content. The practical distinction is timing and purpose: use a process skill to clarify decisions at the start; use a task-specific output skill when the work involves a format whose production or review has its own requirements. Neither is a universal improvement for every coding session.
Free tools Windows power users keep installed
One-click scans. No signup required.
How the two skill roles differ
| Role | Example | When it is used | What it guides |
|---|---|---|---|
| Process | Superpowers brainstorming, followed after design approval by writing-plans | Before implementation | Clarifying decisions, agreeing on a design, and creating a plan with verification steps |
| Output | Anthropic document skills, including DOCX and PowerPoint examples | When producing or editing the deliverable | Format-aware artifact work, including a slide-layout review in the PowerPoint example |
Skills are reusable guidance, not enforcement
Anthropic describes skills as one of several ways to steer Claude Code. A skill exposes its name and description at session start, and its full instructions load when it is invoked. That makes a skill reusable task guidance; it does not make it a hard rule that guarantees a particular result. See Anthropic’s explanation of skills and other Claude Code steering mechanisms.
A skill is a folder containing a SKILL.md instruction file, so its instructions can be adapted to fit a workflow. Anthropic’s current cheatsheet documents project skills at .claude/skills/<name>/SKILL.md and identifies .claude/commands/ as a legacy path. Consult the Claude Code cheatsheet for the documented path and current usage details.
When this pairing is worth trying
- Try the process-first sequence when the request leaves meaningful choices open, or when previous sessions have gone in a direction you did not intend.
- Add an output skill when a deliverable such as a DOCX file or PowerPoint presentation benefits from a repeatable, format-specific procedure.
- Keep instructions scoped to the decisions or output they are meant to guide, and revise them when they do not fit your actual workflow.
- Judge the result in context. Jonker’s account reports a personal benefit but supplies no completion-rate figure or controlled comparison.
The original first-person account, credited to Nolen Jonker and XDA Developers and dated September 20, 2026, is available at World Programming Systems.
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
