Stop prompting Claude to "write docs for this" and getting generic, superficial output that misses half the behavior. Doxaflow runs a structured intake process before writing anything — it knows your audience, your style guide, your output format, and your version. It reads code as the source of truth, flags conflicts when your notes say one thing and the implementation does another, and marks every information gap inline rather than papering over it. It even scans for credentials and API keys, giving you the choice to redact, keep, or remove them before they end up in your docs. The result isn't a plausible-looking document — it's an accurate one that's actually useful for the person reading it.

What it does

Stop prompting Claude to "write docs for this" and getting generic, superficial output that misses half the behavior. Doxaflow runs a structured intake process before writing anything — it knows your audience, your style guide, your output format, and your version. It reads code as the source of truth, flags conflicts when your notes say one thing and the implementation does another, and marks every information gap inline rather than papering over it. It even scans for credentials and API keys, giving you the choice to redact, keep, or remove them before they end up in your docs. The result isn't a plausible-looking document — it's an accurate one that's actually useful for the person reading it.

How it works

Doxaflow works through a structured five-step process. It starts by asking a short series of multiple-choice questions — audience, version, style guide, output format — before writing a single word. It then reads your source material and produces an interpretation summary for your review, flagging gaps, conflicts, and anything sensitive before proceeding. Once confirmed, it writes the documentation using the appropriate structure for your artifact type and applies your chosen style guide consistently throughout. Finally, it renders the output directly to your chosen format — Word, PDF, HTML, Confluence, README, or Markdown — and offers to generate an FAQ, save a reusable template, or reformat for another platform before wrapping up.

Why use this skill

• Guided intake process — every session opens with a structured Q&A that captures audience, version, style guide, language, and output format before a single word is written
• Six output formats in one skill — delivers Markdown, Word (.docx), PDF, HTML, README, and Confluence wiki markup natively, with no external tools required
• Gap detection, not gap filling — wherever source material is incomplete, Doxaflow inserts an inline callout marking exactly what's missing instead of inventing content to fill the hole
• Four built-in style guides — enforces Chicago, AP, Microsoft, or Google style consistently across voice, punctuation, heading case, and number formatting; or supply your own via file upload or URL
• Intelligent update mode — infers all settings from an existing document, shows a structured diff of what will change before writing begins, and preserves everything that isn't being touched
• Built-in security scanning — detects prompt injection attempts in uploaded files and flags credentials, API keys, and PII for user review before they appear in the output