safe-docx GEMINI.md

A local MCP server for reading and editing Word .docx documents while preserving formatting and tracked changes. Tracked changes are the insertions, deletions, and other edits recorded for review in a document.

In plain words
What is it for?
Use it to read, search, inspect, and batch-edit .docx files, including checking for tracked changes. The files stay on the local machine.
Why use it?
It lets an AI edit documents without losing their layout or review history, and provides stable paragraph references for precise changes.

Instructions file for Gemini CLI

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/usejunior/safe-docx/gemini-md
Clone the repo
git clone --depth 1 https://github.com/UseJunior/safe-docx

Made for: Gemini CLI.

Per session 1,009 This file is loaded in full into every session.
When invoked 1,009 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.01009 $0.01009
Opus 5 $0.00504 $0.00504
Sonnet 5 $0.00202 $0.00202
Haiku 4.5 $0.00101 $0.00101

Measured 2d ago against content hash 6ff019bcc598, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

safe-docx GEMINI.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

GEMINI.md · 81 lines

How it starts

The opening of the file, as written. The whole thing — 81 lines — stays where its author put it; the contents beside it link to each section on GitHub.

SafeDocX — AI Document Editing (MCP Server)

SafeDocX is a local MCP server for editing .docx files with full formatting preservation and tracked changes. It runs via npx -y @usejunior/safe-docx using stdio transport.

Trust Boundary

SafeDocX runs locally only — no data leaves the machine. All document reads and writes happen on the local filesystem. There is no remote endpoint.

Available Tools

Reading and Navigation

  • read_file — Read document content with stable paragraph IDs (_bk_*). Supports toon, json, and simple output formats. Use offset/limit for pagination.
  • grep — Regex search across paragraphs. Returns paragraph anchors with match context. Use dedupe_by_paragraph (default true) to get one result per paragraph.
  • get_session_status — Get session metadata including edit count and normalization stats.
  • has_tracked_changes — Check whether the document contains tracked-change markers (insertions, deletions, moves, property changes). Read-only.

Batch Editing

  • batch_edit — Validate and apply multiple edit steps (replace_text, insert_paragraph) in one call. Validates all steps and rejects conflicts (duplicate ids, overlapping replace ranges, insert-slot collisions) before applying anything; an execution-time failure stops at the first failing step and reports which steps completed. Accepts inline steps or a plan_file_path.

Editing

  • replace_text — Find-and-replace within a single paragraph by _bk_* ID. Preserves formatting across run boundaries. Supports inline tags: <b>, <i>, <u>, <highlighting>.
  • insert_paragraph — Insert a new paragraph before or after an anchor paragraph by _bk_* ID. Optional style_source_id to clone formatting from a different paragraph.
  • add_comment — Add comments or threaded replies anchored to paragraphs.
  • get_comments — Get all comments with IDs, authors, dates, anchored paragraphs, and threaded replies. Read-only.
  • delete_comment — Delete a comment and all its threaded replies.
  • accept_changes — Accept all tracked changes in the document body, producing a clean document.

Read the full file on GitHub · 81 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 81 lines · 1,009 tokens per session scan A 6ff019bcc598

Subscribe to this mod's changes

safe-docx GEMINI.md is an instructions file published in the GitHub repository UseJunior/safe-docx (40 stars, last pushed 4d ago), licensed Apache-2.0. It adds 1,009 tokens to every session, about $0.0050 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.