From writing
Reviews documentation, blog posts, READMEs, proposals, and prose for quality issues in content, style, and embedded artifacts using parallel agents.
How this skill is triggered — by the user, by Claude, or both
Slash command
/writing:reviewThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Review a document through parallel agents, each covering a different lens.
Review a document through parallel agents, each covering a different lens.
$ARGUMENTS
If no document path provided, ask the user. Optionally accept target audience and focus areas.
Read the document. Always spawn writing:content and writing:style. Spawn writing:artifacts only if the document contains URLs, markdown links, Mermaid code blocks, or markdown tables.
Spawn all applicable agents in parallel. Each receives the document content, audience, and focus areas.
For writing:style, also inject these writing preferences (sub-agents cannot see skills):
meticulous/meticulously, pivotal, testament, underscore (figurative), interplay, intricacies, bolstered, garner/garnered, foster/fosteringboasts, vibrant, showcasing, nestled, groundbreaking, renowned, diverse arrayMerge agent findings per references/synthesis.md. Deduplicate, rank by severity, and present a unified report.
npx claudepluginhub bendrucker/claude --plugin writingPerforms hierarchical bottom-up document review, diagnosing structural problems and producing REVIEW.md for revision. Called internally after claim validation.
Orchestrates six sub-skills to review documentation files, directories, or PRs: style, outline, UX, code verification, coherence, and pitfalls.
Reviews documentation for data inconsistencies, broken references, typos, and unclear terminology. Use before publishing or committing to catch cross-file quality issues.