Doc Co-Authoring

Turn rough ideas into reader-ready docs without the usual back-and-forth. The Doc Co-Authoring skill guides you through a structured three-stage workflow: Context Gathering (you dump everything you know while Claude asks clarifying questions), Refinement & Structure (build each section through brainstorming, curation, and surgical edits), and Reader Testing (a fresh Claude with no context reads the doc and surfaces blind spots). The result is documentation that works when others read it—including when they paste it into Claude. You get this outcome without hand-holding. Trigger it when you say "write a doc", "draft a proposal", or "create a spec". Claude offers the workflow, you accept, and it walks you through meta-context (doc type, audience, impact), then encourages an info dump—stream of consciousness, links to channels, shared docs. Once context is sufficient, you move section by section: clarifying questions, 5–20 brainstormed options, you curate what to keep/remove/combine, then draft and refine with str_replace so edits are precise. When the draft is done, Reader Testing runs: a sub-agent (or a fresh claude.ai conversation) gets only the document and predicted reader questions. Anything it gets wrong or finds ambiguous gets fixed before you call it done. If you have Slack, Teams, or Drive connected, the skill can pull context from those integrations. If not, it suggests enabling connectors or pasting content. The skill is built for PRDs, design docs, decision docs, RFCs—any structured content where quality and clarity matter more than speed. One install, and every doc you co-author with Claude can follow the same rigor.