Create a chart from a range. Chart types: ColumnClustered, ColumnStacked, BarClustered, BarStacked, Line, LineMarkers, Pie, Doughnut, Area, XYScatterLines. Returns the chart name and title.

Add a threaded comment to a cell.

Add a conditional formatting rule to a range. Supports cellValue (compare with a formula and fill on match), colorScale (min/max gradient), dataBar, and iconSet rules.

Draw a line or curved connector between two shapes or points, with all OOXML bounding-box math (off/ext/flips, bezier path) computed for you. For bezier curves the path leaves and arrives horizontally — no flip attributes are used, avoiding the custGeom double-mirror pitfall. The shape is appended on top of the slide.

Define a named range (named formula) in the workbook, e.g. name "TaxRate" for "=Sheet1!$B$1".

Add a new slide from a layout at a specific position, with placeholder text pre-filled and shapes renamed to layout names. Use inspect_layouts to find available layout names and placeholder names. Returns the new slide number (1-based) and its placeholder shapes with semantic names.

Create an Excel table (ListObject) over a range. Returns the table name and range.

Add a new worksheet to the workbook, optionally at a specific position, and activate it. Returns the new sheet name and position.

Autofit column widths and/or row heights of a range to their content.

Force Excel to recalculate the workbook. Use after writes when you need formula results to be current before reading them back.

Clear the contents and formatting of a range in the live workbook.

Add a comment to a paragraph (anchored on its text).

Copy a source range to a target range in the live workbook. Copies values, formulas and formatting by default; valuesOnly copies just the results.

Copies slides between two open presentations entirely server-side — the Base64 data never enters Claude context. Exports from the source presentation and inserts into the destination in a single operation. Both presentations must be connected to the bridge. Requires PowerPointApi 1.8.

Delete one paragraph (text and paragraph mark) from the document.

Delete a worksheet from the workbook. This removes its contents permanently and can break formulas referencing it.

Duplicate a slide within the same presentation. Exports the slide and reimports it at the specified position. Data stays server-side. The reimport makes PowerPoint view jump to the first slide.

Replace paragraph content of a shape with raw OOXML <a:p> XML. Preserves <a:bodyPr> and <a:lstStyle>. Use read_shape_paragraphs first to get the current XML, modify it (using /pptx skill knowledge), then write it back. The slide is exported, modified server-side, and reimported — data never enters Claude's context. NOTE: the reimport deletes and reinserts the slide, but the view is restored to the edited slide afterwards. For text-only changes, execute_officejs still avoids the reimport entirely (in place via shape.textFrame.textRange).

Create a chart on a slide from structured data. Generates all OOXML automatically (chart XML, rels, graphic frame, Content_Types). Supports column, bar, line, pie, doughnut, and area charts with multiple series.

Edit slide XML and reimport. Two modes: (1) xml mode — provide finished XML string (use read_slide_xml first to get current XML), (2) code mode — provide JS code that manipulates the pre-parsed DOM (receives: doc, findShapeById, NS_P, NS_A, escapeXml, serializeXml, DOMParser). Code mode preserves untouched attributes. The slide is exported, modified server-side, and reimported — data never enters Claude's context. NOTE: the reimport deletes and reinserts the slide, but the view is restored to the edited slide afterwards — reserve this for structural changes (new shapes, layout rebuilds). BULK EDITS: apply ALL edits across slides first, then verify with a single screenshot — do NOT alternate screenshot and edit_slide_xml calls; rapid reimport cycles crash PowerPoint for Mac (reimports are throttled server-side, so interleaving only wastes time).

Update multiple files in the slide zip and reimport in a single operation. Accepts a map of { path: content } — can modify existing files or add new ones (e.g. chart XML + rels). Auto-registers Content_Types for new chart files. Use read_slide_zip first to get the current content.

Set speaker notes on one or more slides. Accepts markdown text (paragraphs, bold, italic). Converts to OOXML server-side and injects via single-slide reimport. Replaces all existing notes on the specified slides.

Add or delete columns of a table. Add: position "Before" inserts left of the first column, "After" right of the last column; optionally pre-fill with values (2D array, one column per entry). Delete: remove count columns starting at columnIndex (0-based). Note: some Word builds reject column operations with InvalidArgument.

Execute arbitrary Office.js code inside the live Excel workbook. The code runs inside Excel.run(async (context) => { ... }) with 'context' available as a variable. Use 'await context.sync()' after loading properties. Return a value to get it back as the tool result. Common operations: read/write ranges (workbook.worksheets), formulas, tables, charts. Cell references use A1 notation; values/formulas are returned as arrays matching the range shape.

Execute arbitrary Office.js code inside the live PowerPoint presentation. The code runs inside PowerPoint.run(async (context) => { ... }) with 'context' available as a variable. Use 'await context.sync()' after loading properties. Return a value to get it back as the tool result. For positioning, all values are in points (1 point = 1/72 inch). Common operations: add shapes, set text, change colors, add/delete slides. Edits run in place without reimporting the slide, so PowerPoint's current view position is preserved — prefer this over edit_shape_paragraphs / edit_slide_xml for text-only or formatting-only changes. Host limitations (live-tested on PowerPoint for Mac): (1) slide.shapes.getItemById and getItemByName do NOT exist — load shapes.load("id"), sync, then iterate shapes.items and match String(sh.id). (2) Loading textFrame paths (textFrame/hasText, textFrame/textRange/text) on shapes without text frames (Graphic/image shapes) fails the whole context.sync() with a bare InvalidArgument — load "id" first, then load textFrame paths only on shapes known to be text shapes, or wrap per-shape loads in separate syncs with try/catch.

Execute arbitrary Office.js code inside the live Word document. The code runs inside Word.run(async (context) => { ... }) with 'context' available as a variable. Use 'await context.sync()' after loading properties. Return a value to get it back as the tool result. Common operations: read the body (context.document.body), insert paragraphs, search and replace, work with tables, read tracked changes.

Find text in a range or the whole used range. Returns the address of the first match, or all matches as a (possibly multi-area) address string.

Format one paragraph: style, alignment, and font (name, size, bold, italic, color, highlight color). Only provided properties are changed.

Format a range: font (name, size, bold, italic, underline, color), fill color, borders (all/outline with style, weight, color), alignment, wrap, and number format. Only the provided properties are changed.

Apply formatting to multiple shapes on a slide in one call. Generates Office.js code internally. Use for fill color, font bold/italic/size/color/name. Cannot set corner radius or borders (use edit_slide_xml code mode for those). Runs in place (no slide reimport), so PowerPoint view position is preserved.

Freeze rows and columns of a worksheet at a cell (everything above and left of the cell stays visible), or unfreeze. E.g. freezeAt "A2" freezes the first row, "B2" freezes row 1 and column A.

Read the document metadata: title, author, subject, keywords, comments, category, creation and last-save dates.

Returns a local file path for the presentation. For local files, returns the existing path. For SharePoint/cloud files, exports server-side and saves to a temp .pptx. Caches by revision number — re-exports only when the presentation has been saved since last export.

Read one paragraph in full: text, style, alignment and font properties (name, size, bold, italic, color).

Read the current user selection (cursor) in the document.

List tracked changes (revisions) in the document: type (insertion/deletion), author, date and a text snippet. Read-only; Office.js cannot create tracked changes.

Inserts an image onto a slide using Office.js setSelectedDataAsync with CoercionType.Image. Accepts a file path, URL, or raw base64 data. Optionally target a specific slide (the view stays there so the user sees the insert; call show_slide first per the view etiquette) and control position/size in points.

Insert or delete a range, shifting existing cells. For insert, shift is Down or Right; for delete, Up or Left.

Insert a page break at the end of the document body.

Insert a new paragraph into the document, optionally with a style (e.g. "Heading 1", "Title", "Quote"). Insert at the start or end of the body, or before/after a given paragraph index. Returns the new paragraph count.

Insert a table at the end of the document, optionally pre-filled from a 2D array of strings. Returns row and column counts.

Deck overview: slide dimensions, theme (colors + fonts), and all slides with slide number (1-based), ID, and shape count. Use as the first call to understand deck structure. Theme is cached after the first call. For shape details, follow up with scan_slide or inspect_slide on specific slides. For available layouts, use inspect_layouts.

Overview of the live document: paragraph count, section count, table count, inline picture count, and a preview of the first paragraphs with their styles. Use this first to orient yourself in an unfamiliar document.

Returns slide layouts with names, OOXML type (e.g. blank, twoObj, secHead), master/layout indices, and detailed placeholders. Use fields to control which data is returned. By default reads all layouts from every slide master in OOXML (complete list, requires file access — may take a moment on first call for cloud files). Layouts sharing a name across different masters are distinguished by masterIndex. Set usedOnly to return only layouts assigned to existing slides (fast, Office.js only, no file access).

Detailed slide inspector (~80 tokens/shape): returns every shape with text content, positions, sizes, and fill colors, plus slide dimensions. Supports slideRange for multiple slides. For just positions without text/fills, use scan_slide instead.

Overview of the live workbook: all worksheets (name, position, visibility, used-range address), all Excel tables (name, sheet, range), all named ranges, and all PivotTables (name, sheet). Use this first to orient yourself in an unfamiliar workbook.

List all content controls (template fields) in the document: id, title, tag, type, current text and placeholder. Use set_content_control_text with the id to fill them.

Lists all Word documents currently connected to the bridge server. Shows document IDs (file paths for saved files, generated IDs for unsaved) and connection status. Use this to find the documentId to pass to other Word tools when multiple documents are open.

Lists all PowerPoint presentations currently connected to the bridge server. Shows presentation IDs (file paths for saved files, generated IDs for unsaved) and connection status. Use this to find the presentationId to pass to other tools when multiple presentations are open.

Lists all Excel workbooks currently connected to the bridge server. Shows workbook IDs (file paths for saved files, generated IDs for unsaved) and connection status. Use this to find the workbookId to pass to other Excel tools when multiple workbooks are open.

Deck preview: batch overview of all/selected slides with optional thumbnails + text. With images: ~900 tokens/slide; text-only (includeImages=false): ~35 tokens/slide. Use for visual review or content audit. Do NOT use to inspect one slide — use inspect_slide or screenshot_slide instead.

Read all comments in the document body: id, author, text, resolved state, date, and replies. Note: adding replies is not supported by the current Word.js API set.

Lightweight text extractor: returns slide titles and body text as plain strings (~20x smaller than inspect_slide). Use for content review, narrative analysis, or any read-only text task. Supports slideRange and optional speaker notes.

Read paragraphs of the document with their styles, paginated by offset and limit. Returns total count plus the requested slice.

Read values, formulas and number formats from a range of the live workbook. Returns a 2D array (rows x columns) plus the range address and dimensions. Use "usedRange": true to read the whole used range of the sheet.

Read raw OOXML <a:p> paragraphs from a shape's text body. Returns the paragraph XML as a string — preserves all formatting (bold, colors, bullets, etc.) that textRange.text strips. Use with the /pptx skill's OOXML knowledge to understand and modify the XML.

Read the full raw OOXML of a slide, or filter to a specific shape. Returns the slide's ppt/slides/slide1.xml content. Use with the /pptx skill's OOXML knowledge to understand the XML structure.

Read multiple files from the exported slide zip. Returns slide XML, relationships, chart XMLs, and Content_Types. Use this to inspect chart data, rels, or other zip contents beyond what read_slide_xml provides. When no paths specified, auto-discovers all text/XML files in the zip.

Read speaker notes from slides. Uses full-deck export (cached by revision) to parse notes from OOXML — Office.js has no notes API. Returns plain text per slide.

Read all cell texts of a table as a 2D array.

Read all threaded comments in the workbook: content, author, creation date, resolved state, and replies. Use reply_to_comment or resolve_comment with a comment id from this output.

Refresh all PivotTables in the workbook. Use after data changes when pivot summaries must be current.

Find and replace text across a whole worksheet (substring match per cell by default). Returns the number of cells changed. Uses findAll so merged/hidden cells behave like the Excel UI.

Find and replace text throughout the document. Returns the number of replacements made.

Add a reply to an existing threaded comment. Use the comment id from read_comments.

Resolve or reopen a threaded comment. Use the comment id from read_comments.

Save the document. If it has never been saved, Word applies its default save behavior (may prompt).

Lightweight shape scanner (~40 tokens/shape): lists shape IDs, types, and positions on slides, plus slide dimensions. Supports slideRange for multiple slides. For text content and fills, use inspect_slide instead.

Slide screenshot (~1000 tokens): captures one slide as PNG image. Use to visually verify layout after changes. Do NOT loop over all slides — use preview_deck instead. WARNING: renders with whatever font the host webview substitutes when the deck font is not system-installed (e.g. Century Gothic) — substituted fonts have different metrics, so text width, overflow and centering judgments from these screenshots are UNRELIABLE. For font-faithful QA export via PowerPoint itself (AppleScript: save pres in POSIX file "/tmp/x.pdf" as save as PDF, then pdftoppm) and measure geometry with Python (PIL pixel measurement) rather than judging by eye.

Search the document body for text. Returns up to 50 matches with a text snippet around each hit.

Search Microsoft Fluent UI icon library. Returns matching icons with svgBase64 ready for insert_image (sourceType: "base64") — the SVG content is fetched server-side so dead icon URLs can never break the workflow. Prefer mono (_M) variants for professional decks. Retry with synonyms if no good matches (e.g. "innovation" → "lightbulb", "security" → "shield").

Search for text across all slides (or a slide range) in the presentation — like grep for slides. Searches shape text, table cells, and speaker notes. Returns matching shapes with context at the chosen level. Case-insensitive substring by default; supports regex.

Replace the text of a content control (template field) identified by its id from list_content_controls. Optionally update its title and tag.

Set a data validation rule on a range. Supports list, wholeNumber, decimal, textLength, and date rules.

Set column widths and/or row heights for a range, in points. Optionally autofit columns or rows to their content.

Set document metadata (title, author, subject, keywords, comments, category). Only provided fields are changed.

Set a hyperlink on a cell (address, display text, tooltip). The cell text is replaced by textToDisplay when provided. Pass address as null to remove the hyperlink.

Merge or unmerge cells in a range.

Replace the text of one paragraph, keeping the paragraph itself (and optionally its style).

Write text into one table cell (replacing its content).

Set the text of multiple table cells in one call. Rows and columns are 0-based indices into the table.

Protect or unprotect a worksheet. Protection restricts user edits; cells formatted as locked are affected.

Bring a slide into view in the live presentation (navigates the user to that slide). Call this BEFORE editing a slide so the user can watch the changes happen; also useful to walk the user through a deck slide by slide.

Sort a range by one or more columns. Fields are given as zero-based column offsets within the range.

Vertically rescale a band of shapes (e.g. shrink the content of a slide by 30%) in ONE atomic operation. Shapes whose top falls inside [topPt, bottomPt] are remapped to start at newTopPt with factor scaleY. Connector-like shapes (cxnSp, line/connector geometries, single-stroke bezier custGeom) get their height rescaled too so endpoints stay attached; all other shapes keep their height unless listed in scaleHeightShapeIds. Text, images and footers outside the band are untouched.

Update an existing worksheet: rename it, move it to a position, change visibility, set the tab color, or activate it. Only provided properties are changed.

Run programmatic checks on a slide: detect overlapping shapes, out-of-bounds shapes, empty text, tiny shapes, unused placeholders, placeholder drift from layout defaults, and full-bleed background covers. Returns a list of issues found. Uses the same shape data as inspect_slide — no OOXML needed.

Write values and/or formulas into a range of the live workbook. Accepts a 2D array (rows x columns) or a single scalar that fills the whole range. Formulas are set separately from values. Also supports setting the number format. Content in the target range is overwritten.