Writing & Content✨gemini🌐universal
Adeu — Native Track Changes for AI
GEMINI.md
Adeu — Native Track Changes for AI
Quick Start
Install this GEMINI.md
curl -fsSL "https://raw.githubusercontent.com/dealfluence/adeu/main/GEMINI.md" -o GEMINI.md# Adeu — Native Track Changes for AI
Adeu is your document redlining engine. It gives you a safe, token-efficient interface to read and edit `.docx` files, producing native Microsoft Word Track Changes rather than rewriting the file.
## Available Tools
### `read_docx`
Reads a DOCX file and returns its content as CriticMarkup-annotated text:
- `{++inserted++}` — tracked insertion
- `{--deleted--}` — tracked deletion
- `{>>comment<<}` — comment
**Key parameters:**
- `file_path` (required): absolute path to the `.docx` file
- `clean_view=true`: returns the accepted/final text with no markup — use this first to understand context
- `mode="outline"`: returns a heading map only — start here on large documents before reading in full
- `mode="appendix"`: returns defined terms and cross-reference anchors — consult before editing legal docs
- `mode="fields"`: returns a ledger of all form fields (content controls) with CC:<N> ID, tag, alias, type, state, and options
- `fields_offset=N`: 0-indexed pagination offset for mode="fields"
- `page=N`: navigate paginated full-text output
- `search_query="text"`: search document for specific text matches
- `max_matches=N`: cap number of search matches returned (default 20)
- `match_offset=N`: 0-indexed pagination offset for search matches
- `full_paragraph=true`: expand search match snippets to complete paragraphs
- `force=true`: bypass whole-document token budget guard refusal for large files
### `process_document_batch`
Applies a list of edits to a DOCX. Edits apply **sequentially** — each one evaluates against the document state produced by the edits before it, so dependent edits may be chained in one batch (a later edit must target the text as it reads after the earlier edits). By default (`partial=true`), valid edits apply while failing edits return detailed error reports; set `partial=false` for transactional rollback if any edit fails validation.
**Key parameters:**
- `file_path` (required): absolute path to the `.docx` file
- `changes` (required): list of change objects
- `partial=true` (default `true`): non-transactional batch execution — valid edits apply while failing edits return detailed error reports. Pass `partial=false` for transactional rollback if any edit fails validation.
- `ignore_control_locks=true`: permit writes to content-locked controls (`sdtContentLocked`)
- `ignore_document_protection=true`: permit writes to read-only or comments-protected documents
- `allow_untracked_writes=true`: permit writes in fill-in-forms protected documents (where Word suppresses change tracking)
**Change types:**
- `modify`: search-and-replace. `target_text` must uniquely identify the passage. `new_text` supports Markdown headings, bold, italic, and `\n\n` for paragraph breaks. Empty `new_text` deletes the passage.
- `set_field`: fill a form field (content control) by `field` ("CC:<N>", tag, or alias) and `value`. Automatically dual-writes bound data stores.
- `accept` / `reject`: finalize or revert a tracked change by `target_id` (e.g. `Chg:12`). Revision ids are numbered per package part; if the same id exists in several parts (e.g. body and a header) the bare id is refused, and the optional `part` field (e.g. `word/header1.xml`) picks the one you mean.
- `reply`: reply to a comment by `target_id` (e.g. `Com:5`)
Always call `read_docx` immediately before any `accept`/`reject`/`reply` — IDs shift between document states.
### `apply_text_revision`
Applies a whole-text revised document against an original DOCX file by computing paragraph and line diffs automatically and converting them to native Track Changes.
**Key parameters:**
- `file_path` (required): absolute path to original `.docx` file
- `revised_text` (required): full text of the revised document
- `author`: optional Track Changes attribution string. Python resolves author from parameter -> ADEU_AUTHOR -> OS user name from getpass.getuser() (machine account names such as root/admin/administrator/system/daemon/nobody excluded) -> "Adeu AI". Node resolves from parameter -> ADEU_AUTHOR -> "Adeu AI (TS)".
### `accept_all_changes`
Accepts every tracked change in one operation, producing a finalized clean document. Use only when review is fully complete.
- `remove_comments` (bool, **default `true`**) — also deletes every comment, because the output is meant to be distributable and comments are internal review notes. Pass `remove_comments=false` to accept the tracked changes while keeping the comments.
- Either way the response reports how many comments were deleted and names each one with its author. A comment whose anchored text an accepted deletion consumes is removed regardless, exactly as Word does.
## Recommended Workflow
1. `read_docx(mode="outline")` — understand document structure
2. `read_docx(clean_view=true)` — read final text for context
3. `read_docx()` — read raw markup to see existing tracked changes and comment IDs
4. `process_document_batch(...)` — apply your editsRepository
168stars
26forks
Details
Related Skills
More writing & content skills you might like
gpt-researcher
Writing & ContentGPT Researcher is an autonomous deep research agent that conducts web and local research, producing detailed reports with citations. Use this skill when helping developers understand, extend, debug, or integrate with GPT Researcher - including adding features, understanding the architecture, working with the API, customizing research workflows, adding new retrievers, integrating MCP data sources, or troubleshooting research pipelines.
✨gemini🤖claude💬gpt🌐universal
27.4k
3.7k
View
xiaohu-wechat-format
Writing & ContentClaude Code 公众号一键排版+发布技能 | Markdown → 微信兼容 HTML → 推送草稿箱 | 30 套主题 + 可视化画廊
🤖claude🌐universal
351
42
View
twitter-algorithm-optimizer
Writing & ContentAnalyze and optimize tweets for maximum reach using Twitter's open-source algorithm insights. Rewrite and edit user tweets to improve engagement and visibility based on how the recommendation system ranks content.
✨gemini🤖claude💬gpt⌨️cursor
77.0k
9.1k
View
yara-rule-authoring
Writing & ContentGuides authoring of high-quality YARA-X detection rules for malware identification. Use when writing, reviewing, or optimizing YARA rules. Covers naming conventions, string selection, performance optimization, migration from legacy YARA, and false positive reduction. Triggers on: YARA, YARA-X, malware detection, threat hunting, IOC, signature, crx module, dex module.
🌐universal
7.5k
643
View