Architecture refine
Skill yarlson/yarstack/plugins/yarstack/skills/architecture-refine
Engineering workflows and standards for Codex and Claude Code: plan, implement, test, review, and deliver repository changes.
npx -y skills add yarlson/yarstack --skill architecture-refineAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Refine underdefined product or system architecture through focused interactive decisions. Use when the user wants architecture sparring before a formal implementation plan or contract update.
SKILL.md
1.6 KB, as published. Nobody here has run it
Architecture Sparring
Resolve only the architecture decisions needed for the user's stated next step.
Interaction Loop
- Inspect the relevant repository evidence and known constraints.
- Identify the highest-leverage unresolved architecture decision.
- Ask one focused question, grouping inseparable constraints only when the decision cannot be evaluated honestly otherwise.
- Offer two or three concrete options with material tradeoffs.
- Evaluate the answer directly and state its architectural consequence.
- Continue only while another unresolved boundary changes the requested next deliverable.
Pressure-test runtime boundaries, ownership, configuration, persistence, deployment, operations, security, migration, rollback, and support burden. Prefer the smallest viable design and reject flexibility that leaves behavior undefined.
Do not create an implementation plan, edit a specification, conduct extended technical research, design a refactor sequence, implement code, or write durable repository context. Use plan-create, spec-update, technical-spike, or repo-context-document for those outputs.
Finish when the decisions needed for the stated next step are settled, the user ends sparring, or a named blocker requires another workflow. Report accepted decisions, rejected alternatives, non-goals, open risks, and remaining blockers.