Refine an artifact through five structured passes: Draft, Correctness, Clarity, Edge Cases, and Excellence.
npx skills add https://github.com/naramore/dotai --skill rule-of-five
Please help me install this Agent Skill. Skill: rule-of-five Source: https://github.com/naramore/dotai/tree/main/claude/skills/rule-of-five Command: npx skills add https://github.com/naramore/dotai --skill rule-of-five
Iterative refinement helps teams produce higher-quality outputs by methodically improving a draft through focused passes.
Start with a draft of your artifact, then perform four single-lens refinement passes—Correctness, Clarity, Edge Cases, and Excellence—on the same work.
High-intent search queries and answers about installing and using this skill.
Iterative refinement is performed through a structured five-pass workflow that applies Correctness, Clarity, Edge Cases, and Excellence passes to a draft, systematically improving a single artifact while preserving context.
Multi-pass refinement is a staged scrutiny process that improves code quality by applying focused review lenses—Correctness, Clarity, Edge Cases, and Excellence—sequentially to the same artifact without re-litigating earlier decisions.
Yes, a structured peer-review process works for technical design docs and RFCs by applying staged refinement passes that evaluate Correctness, Clarity, Edge Cases, and Excellence to elevate non-trivial documentation quality.
The best way to review edge cases without losing context is using a guided review rhythm with a dedicated Edge Cases pass, ensuring systematic improvements are made while preserving the artifact's surrounding context.
No, this refinement workflow requires no external dependencies or specific code-review tools, as it operates as a standalone multi-pass process applicable directly to code, prose, designs, and RFCs.
archubbuckGuide teams through staged documentation co-authoring with prompts and draft artifacts.
kuro0107Generate delivery inventory and _meta.md distribution metadata for a PUBLISH_V version.
juspayAudit code for logic errors, silent failures, and unjustified fallbacks.
phoenixvcRun formatting, linting, type checking, and unit tests across a repository.
