Stats
Actions
Tags
From rosetta
Documents the authentication layer of a project — sessions, tokens, middleware, guards, decorators, the permission model. Use when the user says "document auth", "write docs for authentication", "document the login flow", "document how sessions work", or similar. Runs the write-docs engine with an auth-specific playbook, and asks the user at the start whether to run inline (pause for clarifications) or in background (best-judgment, report when done).
How this skill is triggered — by the user, by Claude, or both
Slash command
/rosetta:doc-authThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Documents the authentication layer of the project — flow, sessions or tokens, middleware / guards, decorators, the permission model. A thin preset on top of `write-docs`, pre-framed for auth topics and loaded with an auth-specific playbook.
Documents the authentication layer of the project — flow, sessions or tokens, middleware / guards, decorators, the permission model. A thin preset on top of write-docs, pre-framed for auth topics and loaded with an auth-specific playbook.
Pre-framed topic: "authentication in this project — flow, session or token handling, middleware/guards, and the permission model".
Playbook to read: ${CLAUDE_SKILL_DIR}/../write-docs/references/auth.md.
Run the same rosetta-docs/ check as write-docs Step 1. If rosetta-docs/src/content/docs is missing, redirect the user to /rosetta:init-docs and stop.
Prompt the user once, verbatim:
Should I run inline or in background?
- inline (default): I'll work step-by-step and pause to ask you clarifying questions when the code is ambiguous (multiple auth schemes, a migration in flight, etc.). You'll see the draft before I write anything.
- background: I'll make best-judgment decisions without pausing and write the page unattended. Faster when the codebase is obvious, weaker when there's ambiguity.
(Type
inline/background, or press enter for inline.)
Record the answer. It drives Step 4.
Read ${CLAUDE_SKILL_DIR}/../write-docs/references/auth.md for the topic-specific guidance. Then follow write-docs Step 2 (detect package manager), Step 3 (read rosetta-docs/agent-docs-rules.md), and Step 4 (read rosetta-docs/src/content.config.ts).
inline: continue with write-docs Step 5 (classify via Diátaxis §4) through Step 12 (report). Apply the auth playbook's rules on where to look, which components fit, and which questions to ask when ambiguous. Honor the "interactive" intent: if the code is unclear, stop and ask before drafting.
background: compose a self-contained prompt that embeds (a) this skill's workflow, (b) the auth playbook, (c) the user's pre-framed topic plus $ARGUMENTS as extra context, and (d) the instruction to make best-judgment decisions without pausing. Dispatch it — either via a forked subagent if your session exposes one, or via a backgrounded claude -p --plugin-dir <plugin-path> "..." through Bash. Return the identifier (subagent ID or PID + log path) so the user can follow along, then stop. Do not run the work yourself.
Use write-docs's Step 12 report format. Cite the auth.md playbook sections that shaped non-obvious choices ("per auth.md: declined <CodeTabs> since the project is single-language").
pnpm -C rosetta-docs check before reporting success.npx claudepluginhub marincervinschi/rosetta-plugin --plugin rosettaGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.