writing-guidelines

An agent skill by vercel-labs, from vercel-labs/agent-skills. Tags: copy-editing, debugging, design, documents, office.

What it does

Review docs/prose for Writing Guidelines compliance. Use when asked to "review my docs", "check writing style", "audit prose", "review docs voice and tone", or "check this page against the writing handbook".

Install

With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:

npx skills add vercel-labs/agent-skills --skill writing-guidelines

Or copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):

git clone --depth 1 https://github.com/vercel-labs/agent-skills
cp -r agent-skills/skills/writing-guidelines ~/.claude/skills/writing-guidelines

Safety box score

Grade B, low risk: 18 of 100 risk points, rated 2026-10-21. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.

This skill is a documentation review helper with no local code execution or file mutation, but it depends on a live remote guidelines file every run. The main risks are network access to GitHub and trusting mutable remote instructions for review behavior and output format.

Category levels, from 0 (none) to 3 (high)
Category	Level	Why
Code & shell execution	0 (none)	No shell commands, code evaluation, or shipped native scripts are present; the skill only describes reading files and using WebFetch.
Network access	1 (low)	It always downloads rules from `https://raw.githubusercontent.com/vercel-labs/writing-guidelines/main/command.md`, which makes an external network request on each review.
Filesystem reach	0 (none)	The skill only says to read specified files and output findings; it does not write, delete, or access paths outside the review target.
Secrets & credentials	0 (none)	It does not ask for tokens, keys, passwords, `.env` files, browser sessions, or other secrets.
Privilege & persistence	0 (none)	There is no sudo, global install, service registration, or OS/shell configuration change.
Instruction hijack surface	2 (moderate)	It tells the agent to follow a remote policy source: `The fetched content contains all the rules and output format instructions`, so a changed upstream file can inject new behavior.
Irreversible actions	0 (none)	The skill does not send messages, deploy, publish, pay, book, or otherwise act outside the review task.
Transparency	1 (low)	The effective rules are not shipped with the skill; they live at a mutable external URL, so behavior cannot be fully reviewed from the repo alone.

Before installing

Open the referenced `command.md` URL and decide whether you trust a mutable remote policy source.
If you need deterministic behavior, pin or mirror the guidelines instead of relying on a live GitHub raw URL.
Run the skill once on a harmless sample doc and confirm it only reads the files you expect.

Rated 2026-10-21 by an automated review of 1 file from the skill's repository.

Source

Repository: vercel-labs/agent-skills (all skills from this repository)
Path: skills/writing-guidelines/SKILL.md
Branch: main
Updated: 2026-09-19

grammar-check — Identify grammar, logical, and flow errors in text and suggest targeted fixes without rewriting the entire text.
google-docs — Author and edit Google Docs with the google-docs tools - structured writing, character and paragraph styling, lists, tables, images, headers/footers.
create-prd — Create a Product Requirements Document using a comprehensive 8-section template covering problem, objectives, segments, value propositions, solution.
docx-creator — Produce Word (.docx) and export existing Word/WPS manuscripts to PDF (Word 转 PDF / 试读版 / 排版修复), preserving revisions.
draft-nda — Draft a detailed Non-Disclosure Agreement between two parties covering information types, jurisdiction, and clauses needing legal review.
humanizer — Rewrite AI-sounding text so it reads like the writer without changing what it says.