doc-coauthoring
Guide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.
Security Assessment
About doc-coauthoring
doc-coauthoring is a workflow skill that guides users through collaboratively writing structured documents — documentation, proposals, technical specs, decision docs, RFCs, and PRDs. Rather than producing a one-shot draft, it has Claude act as an active guide through three stages, addressing the common failure mode where a document makes sense to its author but leaves readers (and downstream Claude sessions) missing critical context.
The three stages are Context Gathering, where Claude asks meta-questions (doc type, audience, desired impact, template, constraints) and invites the user to info-dump freely before asking 5-10 clarifying questions; Refinement & Structure, where the document is built section by section through brainstorming 5-20 options, curating, drafting, and surgical edits, starting with the section that has the most unknowns; and Reader Testing, where the draft is tested against a fresh Claude with no prior context to surface blind spots. Along the way it can pull context from available integrations (Slack, Teams, Google Drive, SharePoint, or other MCP servers), suggest enabling connectors when none are detected, create an artifact scaffold with section placeholders, and flag images lacking alt-text since those are invisible to readers using Claude.
It targets engineers, PMs, and technical writers who need well-structured internal documents and want a repeatable process for transferring context, iterating, and validating readability. The skill is purely a process guide — it reads documents the user points to and writes drafts, but describes no destructive operations, credential handling, or code execution.
FAQ
What kinds of documents is this for?
Structured content such as technical specs, decision docs, proposals, PRDs, design docs, and RFCs — situations where a clear, reader-tested document matters.
What are the three stages?
Context Gathering (transfer all context and answer clarifying questions), Refinement & Structure (build the doc section by section), and Reader Testing (validate with a fresh, context-free Claude).
Do I have to follow the full workflow?
No. The skill offers the structured workflow up front and works freeform if you decline; you can also add more context or move between stages as needed.
Can it pull in context from my team tools?
If integrations like Slack, Teams, Google Drive, SharePoint, or other MCP servers are available it can read from them; otherwise it suggests enabling connectors or pasting content directly.
Why does it ask about image alt-text?
Because when others use Claude to understand the doc, images without alt-text are invisible to it; the skill offers to generate descriptive alt-text so the doc works for those readers.
Install doc-coauthoring
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
getsentry/skills