# Review Googledoc

> Review a Google Doc in Chrome and add anchored comments using MCP browser automation. Can review an already-open doc OR open a doc by name from Google Drive. Use when the user wants to review, comment on, or get feedback on a Google Doc - for any review type including spelling, grammar, factual accuracy, clarity, structure, potential edits, rhetorical effectiveness, or custom criteria. If no specific criteria given, reviews for ALL types. Use when this capability is needed.

- Skill: `tomevault-io/review-googledoc` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add tomevault-io/review-googledoc`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/review-googledoc/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: tomevault-io (https://skillmd.com/u/tomevault-io)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/tomevault-io/review-googledoc

---


# Google Docs Review Skill

Use this Skill to review Google Docs and add anchored comments directly in the document using MCP browser automation.

## Prerequisites
- Google Doc must be open in a Chrome tab
- Claude in Chrome MCP browser tools must be available

## Workflow

### Step 1: Get Document Context
1. Use `mcp__claude-in-chrome__tabs_context_mcp` to get available tabs
2. Look for a Google Doc tab (URL contains `docs.google.com/document/`)
3. **If user named a specific document:**
   - Check if that doc is already open (match by title in tab or page)
   - If NOT open, open it:
     a. Create new tab: `mcp__claude-in-chrome__tabs_create_mcp`
     b. Navigate to Google Drive search: `mcp__claude-in-chrome__navigate` url="https://drive.google.com/drive/search?q=[doc name]"
     c. Wait for results to load, take screenshot
     d. Find and click the matching document to open it
     e. Wait for doc to load, take screenshot to confirm
4. **If no doc specified and no Google Doc tab found:**
   - Ask user which document to review
5. Take a screenshot to see the current document state
6. Determine review scope:
   - If user specified criteria, use those
   - If user did NOT specify criteria, review for ALL of the following:
     - Spelling/grammar errors
     - Factual accuracy
     - Clarity and readability
     - Technical accuracy
     - Structure and organization
     - Potential edits (wording improvements, conciseness)
     - Rhetorical effectiveness (argument strength, persuasiveness, flow)

### Step 2: Read Document Content
Use one of these approaches to get the document text:
1. **Scroll and screenshot** - Take multiple screenshots while scrolling to read all content visually
2. **Select all and copy** - Ctrl+A, then read the selection

### Step 3: Analyze and Identify Issues
Based on the user's requested review type, identify specific issues to comment on.
For each issue, note:
- The exact text or phrase to find (for anchoring the comment)
- The comment to add (description of issue + suggestion)

### Step 4: Add Comments (for each issue)
Use MCP automation to add anchored comments:

1. Open Find: `mcp__claude-in-chrome__computer` action="key" text="ctrl+f"
2. Type search text: `mcp__claude-in-chrome__computer` action="type" text="[phrase to find]"
3. Find it: `mcp__claude-in-chrome__computer` action="key" text="Return"
4. Take screenshot to see where it is
5. Close Find: `mcp__claude-in-chrome__computer` action="key" text="Escape"
6. Triple-click to select paragraph: `mcp__claude-in-chrome__computer` action="triple_click" coordinate=[x,y]
7. Find Add Comment button: `mcp__claude-in-chrome__find` query="Add comment button"
8. Click it: `mcp__claude-in-chrome__computer` action="left_click" ref="ref_xxx"
9. Type comment: `mcp__claude-in-chrome__computer` action="type" text="[comment text]"
10. Find submit button: `mcp__claude-in-chrome__find` query="Comment submit button"
11. Submit: `mcp__claude-in-chrome__computer` action="left_click" ref="ref_xxx"

### Step 5: Report Results
After processing all issues, report:
- Total issues found
- Comments successfully added
- Any issues that couldn't be commented (e.g., text not found)

## Usage Triggers

Claude will automatically use this Skill when you ask things like:
- "Review my Google Doc" (reviews open doc for ALL criteria)
- "Review 'My Draft Post'" (opens doc by name if not already open)
- "Check the Google Doc for spelling errors" (specific criterion)
- "Add comments to the Google Doc about factual accuracy"
- "Review this doc for clarity"
- "Check the structure and flow"
- "Suggest edits for this doc"
- "Review 'Weekly Update' for rhetorical effectiveness"
- "Comment on the open Google Doc"

## Important Notes
- Comments anchor to paragraphs, not specific words (Google Docs limitation via this method)
- Each comment requires ~10 MCP tool calls, so large reviews take time
- If Find can't locate exact text, try a shorter unique phrase from that paragraph
- The Google Doc must be in a Chrome tab that's part of the MCP tab group

---
> Converted and distributed by [TomeVault](https://tomevault.io/claim/thezvi) — claim your Tome and manage your conversions.
<!-- tomevault:4.0:skill_md:2026-04-11 -->

