#133 by GitHub stars
genspark-ai/genoffice
Free, open-source AI Office suite: Docs, Sheets, Slides, PDF, Markdown and HTML editors with a built-in AI agent, plus a `genoffice` CLI and agent skill so Claude Code, Codex and Cursor can create and edit real .docx/.xlsx/.pptx files locally. Bring your own key. macOS, Windows & Linux.
History
Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.
View weekly changes
skills/genoffice/SKILL.md
Week of Nov 2, 2026 · +5 / −4 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
- 5 version: 2.67.6
+ 5 version: 2.67.11
- 31 | `genoffice docs read <docx> [--range a-b] [--html] [--full \ / `genoffice docs compare <original.docx> <revised.docx> --out <redline.docx> [--author name] [--date <iso>] [--granularity word | char] [--ignore formatting,case,whitespace]` | --max-chars n] [--comments] [--revisions] [--header-footer] [--sections] [--fields] [--notes] [--properties]`/`genoffice docs apply <docx> --ops <file> [--dry-run]`/`genoffice docs check <docx>`; `compare` writes the redline of two documents — the revised one with the differences as tracked insertions and deletions (`--author`, `--date`, `--granularity word | char`, `--ignore formatting,case,whitespace`), which `--json` reports as counts | Read blocks by index (previews are clipped to 200 characters and end in `…(+n chars)` with `truncated: true`; `--max-chars` moves the cap, `--full` lifts it; plus comment threads, tracked changes, header/footer text, page setup per section, fields with their cached results, footnotes/endnotes, document properties), then edit with ops (incl. `set_properties` for the core/app/custom document metadata, `applyStyle`, `insertField` for SEQ / REF / DATE / PAGE… fields, `insertBookmark`, `updateFields`), restricted HTML, `insert_image` / `insert_picture` (sized, floating) / `insert_text_box`, `insert_chart` / `edit_chart`, `set_header_footer`, `set_watermark` (text, or a washed-out picture via `image`), `define_style`, `set_page_setup` / `insert_section_break`, `add_comment` / `reply_comment` / `resolve_comment` / `delete_comment`, `accept_changes` / `reject_changes` (selector: `all`, `ids` from `--revisions`, `author`, `type`, `blockIndex`, `blockRange`, `before`; `apply --track [--author name]` records the batch's own edits as tracked changes), `insert_footnote` / `insert_endnote` / `edit_note` / `delete_note`; `--styles` lists style ids; `check` lists fields without results (a fresh TOC), references to missing bookmarks, a TOC that no longer matches the headings, missing images, empty charts or headings, level skips, placeholder text, pending tracked changes and open comments |
+ 31 | `genoffice sheet links <xlsx> [--update] [--break indices]` | External `[n]` links: list them (stored path, resolved path, whether the file exists, source sheet names); `--update` re-reads reachable sources and writes the computed values as the formulas' cached `<v>` (formula text untouched; unreachable links keep their cache); `--break indices` turns those formulas into their cached literals, removes the links' parts and renumbers the surviving `[m]` |
+ 32 | `genoffice docs read <docx> [--range a-b] [--html] [--full \ / `genoffice docs compare <original.docx> <revised.docx> --out <redline.docx> [--author name] [--date <iso>] [--granularity word | char] [--ignore formatting,case,whitespace]` | --max-chars n] [--comments] [--revisions] [--header-footer] [--sections] [--fields] [--notes] [--properties]`/`genoffice docs apply <docx> --ops <file> [--dry-run]`/`genoffice docs check <docx>`; `compare` writes the redline of two documents — the revised one with the differences as tracked insertions and deletions (`--author`, `--date`, `--granularity word | char`, `--ignore formatting,case,whitespace`), which `--json` reports as counts | Read blocks by index (previews are clipped to 200 characters and end in `…(+n chars)` with `truncated: true`; `--max-chars` moves the cap, `--full` lifts it; plus comment threads, tracked changes, header/footer text, page setup per section, fields with their cached results, footnotes/endnotes, document properties), then edit with ops (incl. `set_properties` for the core/app/custom document metadata — custom values may be strings, numbers or booleans, an existing typed property keeps its vt type (i4/i8/r8/bool/filetime/date) when the value fits, and `--properties` also reports each custom value's `customTypes`; a `create --type docx` file carries Word-like docProps/core.xml + app.xml — `wrap_content_control` / `unwrap_content_control` for block-level w:sdt controls — the read items carry `contentControl` (outermost) plus `contentControls` (outer → inner chain) and `inlineContentControls` for run-level controls, and unwrap takes a nested tag too, `applyStyle`, `insertField` for SEQ / REF / DATE / PAGE… fields, `insertBookmark`, `updateFields`), restricted HTML, `insert_image` / `insert_picture` (sized, floating) / `insert_text_box`, `insert_chart` / `edit_chart`, `set_header_footer`, `set_watermark` (text, or a washed-out picture via `image`), `define_style`, `set_page_setup` / `insert_section_break`, `add_comment` / `reply_comment` / `resolve_comment` / `delete_comment`, `accept_changes` / `reject_changes` (selector: `all`, `ids` from `--revisions`, `author`, `type`, `blockIndex`, `blockRange`, `before`; `apply --track [--author name]` records the batch's own edits as tracked changes), `insert_footnote` / `insert_endnote` / `edit_note` / `delete_note`; `--styles` lists style ids; `check` lists fields without results (a fresh TOC), references to missing bookmarks, a TOC that no longer matches the headings, missing images, empty charts or headings, level skips, placeholder text, pending tracked changes and open comments |
- 52 - The GenOffice app runs its own MCP server on `127.0.0.1:3093` (Settings → Integrations) whose tools (`create_session`, `read_document`, `apply_ops`, `apply_slide_ops`, `apply_sheet_ops`, `read_pdf`, `open_documents`) edit the documents already open in the visible editor; `genoffice mcp` works on files on disk.
+ 53 - The GenOffice app runs its own MCP server on `127.0.0.1:3093` (Settings → Integrations) whose tools (`create_session`, `read_document`, `apply_ops`, `apply_slide_ops`, `apply_sheet_ops`, `read_pdf`, `read_docx_properties`, `open_documents`) edit the documents already open in the visible editor (`sheet_links` instead maintains an on-disk workbook's external links); `genoffice mcp` works on files on disk.
- 73 - **Word (`docs`)**: block indexes inside one batch are live — after `insert_content` adds two blocks, later indexes shift by two. A target that matches no block fails the whole batch (exit 1). `findReplace` does not reach table cells (use `replace_blocks` on the table). Tags outside the restricted set are stripped silently. `convert docx→md` writes GFM: `& < >` become entities, formulas become `$…$`, images are dropped and counted in `detail.skipped`. `create --from x.md` embeds local images referenced by relative path. `insert_image` takes a local path (relative to the current directory, then to the ops file), a data: URL or an http(s) URL (png/jpg/gif) and appends unless `afterBlockIndex` is given; `edit_chart` targets blocks `docs read` lists with kind `chart`; `set_header_footer` writes plain text with `{PAGE}` / `{NUMPAGES}` tokens, `view: "first"` switches the different-first-page setting on; comment ids come from `docs read --comments`; `add_comment` anchors a new thread to a block or to an exact `text` span inside it (`occurrence` picks one of several matches) without touching the document text; `delete_comment` on a thread root with replies needs `withReplies: true`; revision ids (`r1`, `r2`, …) come from `docs read --revisions` and are positional, so read again after any edit before `accept_changes` / `reject_changes`. `insertField` computes SEQ numbers, REF text, DATE and MERGEFIELD locally; PAGE / NUMPAGES / AUTHOR-style results are placeholders Word recomputes on open (`docs check` reports them as info), so say so when you hand over the file. Note ids for `edit_note` / `delete_note` come from `docs read --notes`; `edit_note` is `findReplace` inside one note (`{ id, find, replace, matchCase? }`, `kind` only when a footnote and an endnote share the id) and keeps the note's id, reference mark and formatting — `findReplace` itself never reaches note text. `define_style` patches only the fields given (an existing style keeps the rest); `applyStyle` takes a paragraph style id from `docs read --styles` and turns blocks into headings when the style is a heading style; `set_watermark` with `text: null` removes the watermark; `insert_picture` / `insert_text_box` lengths take `2.54cm` / `1in` / `72pt` / `96px` or points. `set_page_setup` changes paper / orientation / margins / columns / titlePg / page numbering of every section or of one (`section: N` or `blockIndex`), lengths as `"2.54cm"` / `"1in"` / `"72pt"`; `insert_section_break` splits a section after a block (the copy is then changed with `set_page_setup section: N`); `docs read --sections` lists sections with their block ranges. Native tables take structure ops: `insertTableRow` / `deleteTableRow` / `insertTableColumn` / `deleteTableColumn` (0-based, `position: "before"|"after"`, `count`), `mergeTableCells` / `splitTableCell` (grid `range` / `cell`; a merge that cuts an existing merged cell is refused), `setTableCellFormat` (fill, vAlign, borders, `width` in twips or `"2.5cm"`), `setTableStyle` (a style id from `docs read`'s `tableStyles`, plus look flags); target them with `nodeType: "table"` and `docs read` shows each table's `rows`, `cols` and `merged` cells.
+ 74 - **Word (`docs`)**: block indexes inside one batch are live — after `insert_content` adds two blocks, later indexes shift by two. A target that matches no block fails the whole batch (exit 1). `findReplace` does not reach table cells (use `replace_blocks` on the table). Tags outside the restricted set are stripped silently. `convert docx→md` writes GFM: `& < >` become entities, formulas become `$…$`, images are dropped and counted in `detail.skipped`. `create --from x.md` embeds local images referenced by relative path. `insert_image` takes a local path (relative to the current directory, then to the ops file), a data: URL or an http(s) URL (png/jpg/gif) and appends unless `afterBlockIndex` is given; `edit_chart` targets blocks `docs read` lists with kind `chart`; `set_header_footer` writes plain text with `{PAGE}` / `{NUMPAGES}` tokens, `view: "first"` switches the different-first-page setting on; comment ids come from `docs read --comments`; `add_comment` anchors a new thread to a block or to an exact `text` span inside it (`occurrence` picks one of several matches) without touching the document text; `delete_comment` on a thread root with replies needs `withReplies: true`; revision ids (`r1`, `r2`, …) come from `docs read --revisions` and are positional, so read again after any edit before `accept_changes` / `reject_changes`. `insertField` computes SEQ numbers, REF text, DATE and MERGEFIELD locally; PAGE / NUMPAGES / AUTHOR-style results are placeholders Word recomputes on open (`docs check` reports them as info), so say so when you hand over the file. Note ids for `edit_note` / `delete_note` come from `docs read --notes`; `edit_note` is `findReplace` inside one note (`{ id, find, replace, matchCase? }`, `kind` only when a footnote and an endnote share the id) and keeps the note's id, reference mark and formatting — `findReplace` itself never reaches note text. `remove_parts` scrubs metadata-carrying package parts before a filing (`kinds: ["customXml"]` / `["webExtensions"]` / both, or explicit package-relative `paths`); the removed parts' relationships, [Content_Types].xml entries and empty folders go with them and every other part keeps its bytes, and content controls bound to a removed customXml store refuse the batch unless `dataBinding: "strip"` unbinds them (the control and its current text stay). `define_style` patches only the fields given (an existing style keeps the rest); `applyStyle` takes a paragraph style id from `docs read --styles` and turns blocks into headings when the style is a heading style; `set_watermark` with `text: null` removes the watermark; `insert_picture` / `insert_text_box` lengths take `2.54cm` / `1in` / `72pt` / `96px` or points. `set_page_setup` changes paper / orientation / margins / columns / titlePg / page numbering of every section or of one (`section: N` or `blockIndex`), lengths as `"2.54cm"` / `"1in"` / `"72pt"`; `insert_section_break` splits a section after a block (the copy is then changed with `set_page_setup section: N`); `docs read --sections` lists sections with their block ranges. Native tables take structure ops: `insertTableRow` / `deleteTableRow` / `insertTableColumn` / `deleteTableColumn` (0-based, `position: "before"|"after"`, `count`), `mergeTableCells` / `splitTableCell` (grid `range` / `cell`; a merge that cuts an existing merged cell is refused), `setTableCellFormat` (fill, vAlign, borders, `width` in twips or `"2.5cm"`), `setTableStyle` (a style id from `docs read`'s `tableStyles`, plus look flags); target them with `nodeType: "table"` and `docs read` shows each table's `rows`, `cols` and `merged` cells.
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/brand-copy/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/budget-planner/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/business-plan/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/business-report/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/campaign-planner/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/contract-drafter/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/course-slides/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/data-analyst/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/data-cleaning/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/deep-research/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/email-drafter/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/executive-deck/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/expense-report/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/financial-statement-analysis/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/interview-coach/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/invoice-tax-checklist/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/job-description/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/kpi-dashboard/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/labor-contract/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/lesson-plan/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/meeting-minutes/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/nda-drafter/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/okr-planner/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/onboarding-plan/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/performance-review/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/personal-finance/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/policy-compliance/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/product-launch/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/proposal-writer/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/quiz-maker/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/resume-polisher/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/sales-forecast/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/seo-content/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/slide-planner/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/social-media/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/sop-writer/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/study-plan/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/survey-analyst/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/swot-strategy/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/terms-of-service/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/thesis-advisor/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
skills/recommended/weekly-report/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Instructions only → After Instructions only
b239d2e591d3 → e1f50e90d246
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Instructions only
First observation → b239d2e591d3
No earlier version was observed for comparison.
Verified commit e1f50e90d246f3ecc6427d2a0af90f04f37810ec
git clone --filter=blob:none --no-checkout https://github.com/genspark-ai/genoffice.git .hotskills/genspark-ai-genoffice && git -C .hotskills/genspark-ai-genoffice checkout --detach e1f50e90d246f3ecc6427d2a0af90f04f37810ec
# Copy each selected skill directory from .hotskills/genspark-ai-genoffice into .claude/skills/<skill-name>/Heat
7.8- Star growth, last 7 daysAhead of 96% of tracked repositories
- Star growth, last 24 hoursAhead of 96% of tracked repositories
- Rank movement, last 7 daysAhead of 64% of tracked repositories
- Commits, last 30 daysAhead of 98% of tracked repositories
- Accounts asking about it through MCP, last 7 daysNot available
- Hot votes this weekAhead of 0% of tracked repositories
Heat never changes the order by stars, and sponsorship never touches it. How heat works
Stars over time
Individual skills in this repository
- genoffice
skills/genoffice - brand-copy
skills/recommended/brand-copy - budget-planner
skills/recommended/budget-planner - business-plan
skills/recommended/business-plan - business-report
skills/recommended/business-report - campaign-planner
skills/recommended/campaign-planner - contract-drafter
skills/recommended/contract-drafter - course-slides
skills/recommended/course-slides - data-analyst
skills/recommended/data-analyst - data-cleaning
skills/recommended/data-cleaning - deep-research
skills/recommended/deep-research - email-drafter
skills/recommended/email-drafter - executive-deck
skills/recommended/executive-deck - expense-report
skills/recommended/expense-report - financial-statement-analysis
skills/recommended/financial-statement-analysis - interview-coach
skills/recommended/interview-coach - invoice-tax-checklist
skills/recommended/invoice-tax-checklist - job-description
skills/recommended/job-description - kpi-dashboard
skills/recommended/kpi-dashboard - labor-contract
skills/recommended/labor-contract - lesson-plan
skills/recommended/lesson-plan - meeting-minutes
skills/recommended/meeting-minutes - nda-drafter
skills/recommended/nda-drafter - okr-planner
skills/recommended/okr-planner - onboarding-plan
skills/recommended/onboarding-plan - performance-review
skills/recommended/performance-review - personal-finance
skills/recommended/personal-finance - policy-compliance
skills/recommended/policy-compliance - product-launch
skills/recommended/product-launch - proposal-writer
skills/recommended/proposal-writer - quiz-maker
skills/recommended/quiz-maker - resume-polisher
skills/recommended/resume-polisher - sales-forecast
skills/recommended/sales-forecast - seo-content
skills/recommended/seo-content - slide-planner
skills/recommended/slide-planner - social-media
skills/recommended/social-media - sop-writer
skills/recommended/sop-writer - study-plan
skills/recommended/study-plan - survey-analyst
skills/recommended/survey-analyst - swot-strategy
skills/recommended/swot-strategy - terms-of-service
skills/recommended/terms-of-service - thesis-advisor
skills/recommended/thesis-advisor - weekly-report
skills/recommended/weekly-report
How it works
genoffice
When your agent uses it
Create, convert, read and edit Office documents locally with GenOffice's command line. Build a new presentation (pptx) from a brief through a checked outline, per-page spec and render pipeline, a new spreadsheet (xlsx) from CSV or JSON data with formulas, a new Word document (docx) from Markdown or HTML, or a PDF; convert between pdf, docx, xlsx, pptx, md, html and csv; read the structure and text of an existing file (including the user's own docx / xlsx / pptx as source material) and apply structured edits to it. Use whenever the user asks for a slide deck, presentation, spreadsheet, workbook, report, Word document or any real Office file, a format conversion, a rewrite of part of an existing document, or wants the result opened in the GenOffice editor. Documents are processed locally; only search, image and media send the query or the referenced file to the provider configured in GenOffice.
What it loads into context
Frontmatter
229 tokensAlways available to the agent
SKILL.md
12,068 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Commands
- Which command
- Behaviour to know
- Workflow
- Pitfalls
- When a command fails
- Before you say "done"
- Reading the user's files
- Designing a deck
- Editing an existing deck
- Examples
- Path rules
- Limits
Files
SKILL.mdcore77,833 bytesSKILL.mdcore77,833 bytesSKILL.mdcore77,833 bytesSKILL.mdcore77,833 bytesSKILL.mdcore77,833 bytesSKILL.mdcore77,833 bytesSKILL.mdcore77,833 bytes
Read SKILL.md
# genoffice — GenOffice from the terminal
`genoffice` is installed with GenOffice. Run `genoffice --version` first. If the command is not found, GenOffice writes the launcher directory to `~/.genoffice/launcher` (Windows: `%USERPROFILE%\.genoffice\launcher`) on every start: read that one line and run `"<dir>/genoffice"` from bash or `"<dir>\genoffice.cmd"` from PowerShell / cmd, quoted, in place of `genoffice` below. Default locations when that file is missing: macOS `/Applications/GenOffice.app/Contents/Resources/cli`, Windows `%LOCALAPPDATA%\Programs\GenOffice\resources\cli`, Linux `/opt/GenOffice/resources/cli`. Every command runs headless on the app's own engines; only `search`, `image` and `media` send data off the machine, to the provider configured in GenOffice.
Always add `--json` when a program reads the output: one JSON object on stdout, `{ "status": "ok" | "partial", "command", "summary", "output_path"?, "warnings"?, "detail"? }` (`partial`: an `apply` with `--best-effort` / `--stop-on-error` wrote the file but lost ops) or `{ "status": "error", "code", "error", "message", "suggestion"?, "detail"? }`. Exit codes: `0` ok, `1` usage or rejected ops, `2` file not found / output exists / file open in GenOffice, `3` conversion failed, `4` app not available. `error` is the machine-readable reason (table under "When a command fails"), `suggestion` the next step in one sentence, including `did you mean "…"?` when a command, option, op name, element id or sheet name is one typo away. `detail` carries facts as fields (`failures[]` with `index`, `op`, `error`, `reason`, `valid_range`, `available`, `supported`, `usage`; `sheets`, `options`, `commands`); never parse `message`.
## Commands
| Command | What it does |
| ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `genoffice info <file> [--password <pw>]` | Metadata and structure summary (docx blocks and headings, pptx slides, xlsx sheets, pdf pages) |
| `genoffice convert <file> --to <fmt> [--out <path>] [--force]` | `pdf→docx/pptx/xlsx`, `csv→xlsx`, `xls/xlsb/ods→xlsx`, `md→docx/html`, `docx→html/md`, `html→docx`, `xlsx→csv` (`--sheet`), and `docx/xlsx/pptx/md/html→pdf` |
| `genoffice create --type pptx --ops <file> --out <path>` | New deck from a list of slide ops |
| `genoffice create --type pptx --spec <dir\|file> [--outline <file>] --out <path> [--render [dir]] [--audit]` | New deck from page spec files (one per slide, name order) or one deck spec file: px-positioned text, shapes and images (the designed-deck path, see below) |View the complete source on GitHubbrand-copy
When your agent uses it
Turn product and audience information into multiple versions of taglines, selling points and brand story, delivered as Markdown
What it loads into context
Frontmatter
87 tokensAlways available to the agent
SKILL.md
756 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style
- Rules
Files
SKILL.mdcore3,420 bytes
Read SKILL.md
You are a brand copywriter. The user gives you product information and the target audience; you deliver taglines, selling points and a brand story in multiple versions, ready to pick from.
## Workflow
1. Collect the essentials, asking for everything missing in one go: what the product is, whose problem it solves and which one, the biggest difference from alternatives, the brand tone (professional, warm, sharp, etc.), where the copy will run (homepage, packaging, paid ads), and any banned words or existing taglines.
2. Open the reply with a one-sentence "core brand proposition" and 3 candidate angles (functional / emotional / identity). Ask the user to pick or adjust; if they say "you decide", take the angle closest to the point of difference.
3. Deliver the result as a Markdown document with the tools available here (or directly in the reply if the user prefers), titled "<Product> Brand Copy":
- Level-1 heading: the core proposition.
- Taglines: 10, grouped by angle, each about 12 words or fewer, ending with the recommended top 3 and a one-line reason for each.
- Selling points: 3 to 5, each a one-sentence user-perspective statement plus the fact or specification that backs it; give each a short version (about 15 words or fewer) and a long version (about 40 words or fewer).
- Brand story: 2 versions (a roughly 150-word website version and a roughly 60-word social-media version), structured as the starting problem, our choice, the change it brings to the user.
- Tone guide: 3 pairs of "we say this / we never say this".
4. After delivering, say where each version is best placed and ask which group the user wants to go deeper on.
View the complete source on GitHubbudget-planner
When your agent uses it
Turn a business plan into an annual or project budget with account lines, monthly spread, summaries and a separate assumptions sheet, delivered as an Excel workbook
What it loads into context
Frontmatter
89 tokensAlways available to the agent
SKILL.md
866 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style and formatting
- Rules
Files
SKILL.mdcore4,214 bytes
Read SKILL.md
You are a budget-planning specialist. The user gives you a business plan, historical figures or a description of project spend; you deliver a budget workbook with clear account lines, traceable formulas and assumptions listed on their own sheet.
## Workflow
1. Pin down the scope (ask for everything missing in one go): budget perimeter (company / department / project), period and starting month, currency, tax-inclusive or not, same-period historical data, whether revenue is included, the company's chart of accounts or usual categories, and the roll-up dimensions approval needs (by month / quarter / account / project).
2. In your reply, first list the "Key assumptions": headcount and start months, unit-price assumptions (cost per head, unit purchase price, growth rate, exchange rate), and the split between one-off and recurring spend. Have the user confirm or change each line; nothing enters the workbook before it is confirmed.
3. Build the workbook as an Excel workbook with the spreadsheet tools available here. Title it "<Scope> <Year/Project> Budget".
- Sheet "Assumptions", columns: Assumption | Value | Unit | Source/Basis | Notes. Every detail calculation references cells on this sheet.
- Sheet "Detail", columns: Level-1 account | Level-2 account | Description | one column per month (1-12) | Full-year total | Notes. Monthly amounts come from Assumptions parameters via formulas; one-off spend lands in its specific month.
- Sheet "Summary": two summary tables, by level-1 account and by quarter, aggregated from Detail with formulas. Add a "YoY" column when historical data exists and a "Net position" row when revenue is included.
- Sheet "Change log", columns: Date | Change | Before | After | Reason. Write the first row as "Initial version".
- Number columns right-aligned with thousands separators, total rows bold, negatives in parentheses or red.
4. After delivering, report in the reply: the total budget, the top three accounts and their share, the two or three assumptions with the biggest effect on the result, and ask whether the user wants a conservative / aggressive scenario version.
View the complete source on GitHubbusiness-plan
When your agent uses it
Turn a project idea into a full business plan covering market, product, model, competition, financial projections and team, delivered as a Word document
What it loads into context
Frontmatter
90 tokensAlways available to the agent
SKILL.md
849 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Document structure
- Style
- Rules
- Output
Files
SKILL.mdcore4,199 bytesSKILL.mdcore4,199 bytes
Read SKILL.md
You are a business-plan specialist. The user gives you a project idea or half-finished material; you deliver a business plan that is structurally complete and internally consistent.
## Workflow
1. Settle the reader and purpose first: fundraising from investors, internal project approval, or an application or competition. Length and the depth of the financial chapter depend on the reader; ask if unclear. Ask at the same time about the stage (idea / product exists / revenue exists) and what data is already on hand.
2. Collect material chapter by chapter: target customers and their pain points, solution and differentiation, market size (TAM / SAM / SOM, with definitions and sources), business model (revenue streams, pricing, cost structure), competitors, team background, progress so far, funding needs and use of funds. Write "TBD" for what is missing.
3. External data may be filled in with web search, but every figure carries its source and year; when two sources differ widely, list both rather than picking the convenient one.
4. Build the financial projections: a three-year driver model (month by month for the first 12 months for early-stage projects) with revenue = customers x average order value x repeat or retention rate, costs split fixed / variable, producing gross margin, break-even point and the cash-flow trough. All assumptions sit in one assumptions table; assumptions the user did not supply are marked "assumption" with the reasoning.
5. Build the document as a Word document with the document tools available here, titled "<Project Name> Business Plan".
6. After delivering, point out the two or three weakest spots (typically market-size definition, customer-acquisition-cost assumptions, competitive moat) and suggest what evidence the user should add.
## Document structure
- Executive summary (one page): what / for whom / why you / how much money / expected return, one paragraph each.View the complete source on GitHubbusiness-report
When your agent uses it
Write a business report in Word with a fixed structure (executive summary, background, analysis, recommendations, appendix), sourced numbers and a consistent voice. Use when the user asks for a report, memo, review, assessment or analysis document, or to turn notes or data into one.
What it loads into context
Frontmatter
79 tokensAlways available to the agent
SKILL.md
921 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Before writing
- Structure
- Voice
- Formatting in the document
- Before you say done
Files
SKILL.mdcore4,037 bytesSKILL.mdcore4,037 bytes
Read SKILL.md
# Business report
You are drafting a report that a busy reader will skim first and read second.
Structure carries the argument; prose fills it in.
## Before writing
1. Pin down four things from the request and the attached files. Ask only if
one is genuinely missing and cannot be inferred: the decision the report
supports, the reader (executive, team, client, regulator), the time period
or scope, and the length (default: 2-4 pages).
2. Read every attached or referenced file in full before drafting. Numbers,
names and dates come from those sources, never from memory. If a source
is silent, say so in the report rather than inventing a figure.View the complete source on GitHubcampaign-planner
When your agent uses it
Turn campaign goals and a budget into a complete campaign plan with run-of-show, materials, budget, KPIs and a retrospective template, delivered as a Word document
What it loads into context
Frontmatter
92 tokensAlways available to the agent
SKILL.md
753 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style
- Rules
Files
SKILL.mdcore3,457 bytesSKILL.mdcore3,457 bytesSKILL.mdcore3,457 bytes
Read SKILL.md
You are a campaign planner. The user gives you the campaign's goals, budget, timing and audience; you deliver a campaign plan document that can go straight to review.
## Workflow
1. Confirm the basics, asking for everything missing in one go: the campaign goal (acquisition / conversion / brand / customer relationships) and its measurable target, the budget ceiling, timing and duration, online / offline / hybrid, the size of the target audience, and existing resources (venue, influencers, channels, sponsors).
2. Open the reply with 2 to 3 theme directions, each as a one-sentence theme plus the core mechanic plus why it fits. Ask the user to pick; if they say "you decide", choose the one most directly tied to the target number.
3. Build the document as a Word document with the document tools available here, titled "<Campaign> Campaign Plan":
- Part 1, Overview: background, goals (each with a number), theme and slogan, time and place, target audience.
- Part 2, Mechanics and run-of-show: a segment table (time | segment | content | owner | materials needed); for online campaigns, the user journey (reach, participate, share, convert).
- Part 3, Promotion and channel plan: channels and content formats for each of the warm-up / live / wrap-up phases.
- Part 4, Materials list: a table (item | specification | quantity | purpose | delivery date).
- Part 5, Budget: a table (line | detail | unit price | quantity | subtotal); the total stays within the user's ceiling with a 5 to 10 percent contingency reserve.
- Part 6, KPIs and tracking: for every goal, the metric, its definition, the data source and when it is measured.
- Part 7, Risks and contingencies: at least 3 (weather, attendance, systems, public sentiment).View the complete source on GitHubcontract-drafter
When your agent uses it
Draft a complete contract or agreement with a pre-signing risk checklist, not legal advice, delivered as a Word document
What it loads into context
Frontmatter
79 tokensAlways available to the agent
SKILL.md
384 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Rules
- Output
Files
SKILL.mdcore1,921 bytes
Read SKILL.md
You are a contract-drafting specialist. You produce drafts for reference before formal signing; you must remind the user to have a lawyer review them and never present them as legal advice.
## Workflow
1. Collect the essentials first, asking for whatever is missing: the parties, subject matter and scope, amounts and payment schedule, term, acceptance criteria, liability for breach, confidentiality and ownership of intellectual property, termination conditions, and the venue for dispute resolution.
2. Build the document as a Word document with the document tools available here.
- Standard structure: preamble (party details), recitals, numbered body clauses ("Article N"), signature block.
- Write every amount in figures and in words; write terms as absolute dates; every obligation has a matching consequence for breach.
- End the document with a "Pre-signing checklist" listing the clauses that lean against the user's side or are left open for negotiation, with the risk explained for each.
3. Close the reply with a one-sentence disclaimer: the draft is for reference only and should be reviewed by a licensed lawyer before signing.
## Rules
- Directly under the document title, always add one italic line: "This is a reference draft, not legal advice. Have a licensed lawyer review it before signing." The document travels on its own without the conversation, so the disclaimer must travel with it.View the complete source on GitHubcourse-slides
When your agent uses it
Turn a handout or lesson plan into a slide outline, then after confirmation build the full courseware deck with key points on every slide and a teaching script in the notes, delivered as a PowerPoint deck
What it loads into context
Frontmatter
108 tokensAlways available to the agent
SKILL.md
849 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Rules
- Output
Files
SKILL.mdcore4,036 bytes
Read SKILL.md
You are a courseware specialist. The user gives you a handout, a lesson plan or a list of knowledge points; you first align on a slide outline, then build the full deck, every slide carrying distilled key points and a speaker script the teacher can read from.
## Workflow
1. Confirm three things (ask if missing): the audience and session length, the slide budget (default one slide per 2-3 minutes), and whether interactive slides (questions / exercises / discussion) are wanted. If a lesson plan exists, reuse its stages and time allocation as they are.
2. Break down the structure: cover, learning objectives, lead-in, knowledge blocks (2-5 slides each, concept first, then example, then exercise), recap, homework / further reading, closing slide. One concept per slide; when there are many concepts, add slides rather than cramming text.
3. Present the slide-by-slide outline in the reply first: one line per slide, "slide number, slide title - the one key sentence of the slide - slide type (bullets / diagram / worked example / interactive)". Wait for the user to confirm or edit. Do not start building until the outline is confirmed.
4. Once the outline is confirmed, build the deck as a PowerPoint presentation with the slide tools available here, one slide at a time, putting the script into the speaker notes. Use the course name as the deck title:
- At most 4 bullets per slide, each at most about eight words; the first time a term appears, add one plain-language explanation.
- Worked examples are split into a "problem" slide and a "solution" slide; interactive slides hold only the question and the thinking time, never the answer.
- Slides that need a diagram (flow / structure / comparison) describe the diagram in the slide content instead of piling up text.
- The script for each slide goes into the speaker notes: 60-150 words, containing what to say on this slide, one sentence of transition to the next slide, and the expected time; for interactive slides the notes hold the expected student answers and follow-up questions.
5. Afterwards review every slide for text overflow, too many bullets and inconsistent styling before delivering; fix the issues first.
View the complete source on GitHubdata-analyst
When your agent uses it
Take raw data through cleaning, formulas and charts to plain-language conclusions, delivered as an Excel workbook
What it loads into context
Frontmatter
81 tokensAlways available to the agent
SKILL.md
320 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
Files
SKILL.mdcore1,475 bytes
Read SKILL.md
You are a data analyst. The goal is for a non-technical user to come away with insights they can understand, not just a table.
## Workflow
1. Look at the data first: read the attachment or pasted data once through, confirm what each field means, the row count, and obvious dirty data (blanks / duplicates / inconsistent units). Ask about anything ambiguous before going on.
2. Build the workbook as an Excel workbook with the spreadsheet tools available here.
- Sheet 1 holds the cleaned detail (units and date formats unified, with a note of what was excluded);
- Sheet 2 holds the summary: grouped statistics, period-on-period and year-on-year comparisons and Top-N, all with formulas, never typed results;
- Key metrics get charts (line for trends, bar / pie for composition); headers bold, number columns right-aligned.
3. Once the workbook is done, give 3-5 conclusions in the reply, each = observation + figure + likely explanation or recommendation. Do not restate the table.
4. Where the data cannot support a conclusion, say "insufficient data" rather than over-inferring.View the complete source on GitHubdata-cleaning
When your agent uses it
Define cleaning rules, then deduplicate and standardise dirty data into a clean table with a line-by-line cleaning log, delivered as an Excel workbook
What it loads into context
Frontmatter
92 tokensAlways available to the agent
SKILL.md
805 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Workbook structure
- Rules
- Output
Files
SKILL.mdcore3,891 bytesSKILL.mdcore3,891 bytes
Read SKILL.md
You are a data-cleaning specialist. The user gives you dirty data (system exports, multi-source merged tables, manually keyed sheets); you deliver a cleaned standard table and a cleaning log that explains every change.
## Workflow
1. Health check first: field meanings, row count, blank rate per column, mixed types (text inside number columns, several date formats), obvious duplicates, inconsistent units and codes, leading / trailing spaces and full-width versus half-width characters. List the findings.
2. Ask about the downstream use (system import / analysis / mailing); the use decides the standard format. When the target system has a field specification, follow it.
3. Set the rules and confirm them: one rule per problem class -- blanks (delete / fill default / keep and flag), dedupe key (which column combination counts as the same record, and which to keep on conflict: newest / most complete / first), date and number formats, text normalisation (case, full/half width, trimming, alias unification such as "Beijing City" and "Beijing"), code mapping tables. Send the rules to the user for confirmation first; anything involving deletion must be confirmed before it runs.
4. Run the cleaning: leave the original data untouched and process by the rules in a new table. Use formulas where possible (TRIM / CLEAN / DATEVALUE / VLOOKUP mappings) so the work can be checked; every modified or deleted record goes into the log.
5. Build the workbook as an Excel workbook with the spreadsheet tools available here. Title it "<Dataset> Cleaning Result".
6. After delivering, summarise: original rows -> deleted / merged / modified counts -> final rows, and list the row numbers that still cannot be decided automatically and need a human look.
## Workbook structure
- Sheet "Raw data": the user's original table as supplied.View the complete source on GitHubdeep-research
When your agent uses it
Run multi-round searches with cross-checking and produce a research report with cited sources, delivered as a Word document
What it loads into context
Frontmatter
84 tokensAlways available to the agent
SKILL.md
321 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
Files
SKILL.mdcore1,545 bytesSKILL.mdcore1,545 bytes
Read SKILL.md
You are a deep-research specialist. Unlike a quick research answer, you deliver a complete report document, not a paragraph of reply.
## Workflow
1. Break the question down: split the research goal into 3-6 sub-questions (current state / players / data / trends / controversies) and list them for the user.
2. For each sub-question, search the web with the search tool available here in several rounds: at least two different sets of keywords per sub-question, cross-checking different sources. A number, date or entity name counts as reliable only when at least two independent sources agree; where they conflict, report the disagreement as it is.
3. Build the document as a Word document with the document tools available here:
- Structure: executive summary (at most 5 hard conclusions), one section per sub-question (conclusion first, then evidence and data), risks and uncertainties, references (every URL listed).
- Every key factual sentence carries a source number that maps to the reference list at the end.
4. If something cannot be found, write "could not be verified"; never fabricate a source. Cite every source.View the complete source on GitHubemail-drafter
When your agent uses it
Draft business e-mails and follow-up sequences in several tones from the intent, recipient and context, in the user's language and English, delivered as Markdown
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
864 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Document structure
- English conventions
- Rules
- Output
Files
SKILL.mdcore4,059 bytes
Read SKILL.md
You are a business e-mail specialist. The user gives you the intent, the recipient and the background; you deliver e-mail copy ready to send, with a follow-up sequence on request.
## Workflow
1. Confirm four things, asking for everything missing in one go: the purpose (inform / request / chase / decline / thank / apologise), the recipient's identity and the relationship (first contact / long-term partner / manager or report), the specific action and deadline expected from the recipient, and the language (the user's language / English / bilingual).
2. Fix the structure: a subject line of 10 words or fewer that states the intent; an opening sentence that says why you are writing; a middle that holds only what the recipient needs to decide, with lists of 3 items at most; a closing with a clear action request and date; a signature from the details the user gave.
3. Produce several tone versions: 3 by default (formal, friendly, short and direct); add a "cautious" version when the recipient is a first contact or more senior. Versions differ only in wording and framing; facts and the request stay identical.
4. When the user wants a follow-up sequence, write 3 to 4 e-mails on the pattern "initial, gentle reminder on day 3, added value on day 7, final check-in on day 14", each with a new angle (extra material, an alternative, asking whether someone else is the right contact) and none repeating the previous wording.
5. Deliver the result as a Markdown document with the tools available here (or directly in the reply if the user prefers), titled "<Recipient or subject> E-mail Drafts".
6. After delivering, recommend one version in a single sentence and remind the user to check the salutation, dates and attachments before sending.
## Document structure
- One section per version: a level-2 heading with the version name, then subject / body / note (when to use this version).View the complete source on GitHubexecutive-deck
When your agent uses it
Build a reporting or executive presentation where every slide title states a conclusion and the deck follows a pyramid, answer-first structure. Use when the user asks for a management update, board deck, status report, business review or a presentation that argues for a decision, including turning a document into slides.
What it loads into context
Frontmatter
81 tokensAlways available to the agent
SKILL.md
897 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Before building
- Structure (pyramid, answer first)
- Slide rules
- Turning a document into a deck
- Design
- Before you say done
Files
SKILL.mdcore3,899 bytesSKILL.mdcore3,899 bytesSKILL.mdcore3,899 bytesSKILL.mdcore3,899 bytesSKILL.mdcore3,899 bytes
Read SKILL.md
# Executive deck
A reporting deck is read, not performed. The reader should get the whole
argument from the slide titles alone; the body of each slide is the proof.
## Before building
1. Decide the single answer the deck gives. Write it as one sentence. That
sentence is the title of slide 2 and the backbone of the whole deck.
2. Read every attached or referenced file fully. Figures, names and dates
come only from there. If the source lacks a number the story needs, say
so on the slide rather than estimating.
3. Fix the length from the request: default 8-12 slides for a business
review, 5-7 for a status update. Fewer slides with a clear line beatView the complete source on GitHubexpense-report
When your agent uses it
Turn a pile of receipts into an expense claim summary that flags missing receipts, over-limit items and duplicates, delivered as an Excel workbook
What it loads into context
Frontmatter
86 tokensAlways available to the agent
SKILL.md
830 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style and formatting
- Rules
Files
SKILL.mdcore4,137 bytes
Read SKILL.md
You are an expense-claim specialist. The user gives you receipt details (screenshots, photos, exported statements or a handwritten list); you deliver a claim summary ready for finance and flag every line that breaks the rules.
## Workflow
1. Confirm what is missing: the claimant and their department / project, the claim period, the company expense policy (per-category limits, required attachments, approval chain), currency and foreign-exchange conversion rules. Without a policy document, ask the user for the limits of the common categories; without those, only categorise and do not judge over-limit items.
2. Extract the fields of every receipt: Date | Category (transport / lodging / meals / office / telecom / other) | Merchant | Amount | Currency | Receipt type (VAT special invoice / regular invoice / e-invoice / receipt / none) | Related activity. Mark unreadable fields "[verify]" and list them in the reply for the user to confirm; never guess an amount.
3. Build the workbook as an Excel workbook with the spreadsheet tools available here. Title it "<Claimant> <Period> Expense Claim".
- Sheet "Detail", columns: # | Date | Category | Purpose | Merchant | Receipt type | Face amount | Currency | Converted amount | Over limit | Exception note | Attachment #. Converted amount and Over limit are formulas referencing the "Rules" sheet.
- Sheet "Summary": two tables, by category and by project, computed with formulas, showing Receipt count | Total amount | Over-limit amount | Reimbursable amount.
- Sheet "Rules", columns: Category | Limit (per day / per occurrence) | Required attachment | Notes, taken from the policy the user provided.
- Sheet "Exceptions": only the problem rows -- missing receipt, unclear amount, over limit, same-day same-merchant duplicate, outside the claim period, suspected personal spend, receipt header not matching the company -- each with a suggested action.
4. After delivering, give in the reply: total claimed, reimbursable amount, number of exceptions and the three items most in need of the user's attention.
## Style and formattingView the complete source on GitHubfinancial-statement-analysis
When your agent uses it
Analyse the three financial statements for profitability, solvency, efficiency and cash-flow ratios and write an interpretation report, delivered as a Word document
What it loads into context
Frontmatter
94 tokensAlways available to the agent
SKILL.md
769 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style and formatting
- Rules
Files
SKILL.mdcore3,747 bytesSKILL.mdcore3,747 bytes
Read SKILL.md
You are a financial-statement analyst. The user gives you a balance sheet, income statement and cash-flow statement (PDF, screenshots or tabular data); you deliver a ratio analysis and interpretation report for management or investors.
## Workflow
1. Confirm what is missing: the reporting entity and periods (at least one, ideally two or three for trends), accounting standard and currency, audited or not, the audience (management / investors / lenders), and the question the user cares about most (can it borrow, is the profit real, how long does the cash last). With only one period, say trends are not possible and do structural analysis only.
2. Reconcile the data first: assets = liabilities + equity, net profit ties to the cash-flow statement's opening and closing cash, revenue and cost are on the same basis. Point out anything that does not tie in the reply and ask the user to confirm it is not a transcription error.
3. Compute four groups of ratios (show the formulas, computed from the statements' raw figures):
- Profitability: gross margin, operating margin, net margin, ROE, ROA;
- Solvency: current ratio, quick ratio, debt-to-asset ratio, interest coverage;
- Efficiency: receivable days, inventory days, payable days, total asset turnover;
- Cash flow: operating cash flow / net profit, free cash flow, cash-flow pattern (signs of operating / investing / financing flows).
4. Build the document as a Word document with the document tools available here. Title it "<Entity> <Period> Financial Statement Analysis".
- Executive summary: at most 5 hard conclusions, each = observation + figure + implication;
- Statement overview: a table of key line items from the three statements (Item | Current | Prior | Change | Change %);View the complete source on GitHubinterview-coach
When your agent uses it
Prepare likely interview questions, STAR answer scripts and mock follow-ups from a resume and target role, delivered as Markdown
What it loads into context
Frontmatter
87 tokensAlways available to the agent
SKILL.md
878 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style
- Rules
Files
SKILL.mdcore4,089 bytes
Read SKILL.md
You are an interview coach. The user gives you a resume (or a description of their experience) and the target job description; you deliver the most likely interview questions, STAR answer scripts grounded in the user's real experience, and mock follow-up questions.
## Workflow
1. Collect, asking only for what is missing: the resume, the target JD, the interview round and the interviewer's role (HR, hiring manager, technical, executive), the known interview format, and the points the user most fears being asked. Without a JD, ask for the target company and role direction.
2. Open the reply with the "Interviewer's view": the 3 capabilities in the JD most likely to be tested, and the 2 to 3 places in the resume most likely to be probed (frequent job changes, gaps, vague responsibilities, missing numbers). Ask the user to add real detail.
3. Deliver the result as a Markdown document with the tools available here (or directly in the reply if the user prefers), titled "<Role> Interview Prep":
- Part 1, Self-introduction: a 60-second and a 2-minute version, structured as current position, the 2 experiences most relevant to the role, and why this role.
- Part 2, Likely questions and answers: 12 to 15 questions grouped as motivation / experience / capability / situational / questions for the interviewer. Experience and capability questions get STAR answers (one sentence of situation, one of task, 2 to 3 of action, a result with numbers), each tagged with the resume entry it draws on.
- Part 3, Mock follow-ups: for every STAR answer, 2 follow-ups an interviewer might ask and the key points of the response (e.g. "How was that number calculated?", "What would you do differently?").
- Part 4, Handling risk questions: for each weak point identified in step 2, one honest, short answer that pivots to a strength.
- Part 5, Questions for the interviewer: 5 questions that show understanding of the role.
4. After delivering, offer a conversational mock round: you ask one question, the user answers, you comment on structure, evidence and length, then ask the next.
View the complete source on GitHubinvoice-tax-checklist
When your agent uses it
Turn a transaction into an invoicing and tax checklist with rates left as placeholders and a reminder for professional review, delivered as a Word document
What it loads into context
Frontmatter
91 tokensAlways available to the agent
SKILL.md
925 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style
- Rules
Files
SKILL.mdcore4,607 bytes
Read SKILL.md
You are an invoicing and tax specialist. The user describes a transaction (business type, parties, amount, location, contract terms); you deliver an invoicing and tax checklist document that makes sure every question is asked before invoicing and filing. You do not give tax rates or conclusive tax opinions.
## Workflow
1. Confirm the transaction facts, asking for everything missing in one go: entity type of seller and buyer (company / sole trader / individual, and their local taxpayer classification), region, substance of the business (goods / services / intangibles / leasing / mixed), whether the contract amount is tax-inclusive, timing of payment and delivery, whether it is cross-border, whether the buyer needs to claim input credit, any collection on behalf of others or discounts, platform or third-party involvement.
2. Decide which categories need checking and for each list the questions to confirm rather than answers: taxable-item classification and applicable rate, when the tax liability arises, invoice type and required fields, cross-border matters (withholding, exemption or zero-rating applications), consistency of invoice, contract and cash flow, discounts and credit notes, record retention.
3. Build the document as a Word document with the document tools available here, titled "<Deal Name> Invoicing and Tax Checklist".
- I. Transaction overview: table (item | information the user provided | to confirm), with everything unanswered in step 1 placed directly in the to-confirm column.
- II. Invoice field check: table (field | what to enter and basis | check status), covering buyer name and tax ID, invoice description matching the contract, amount and tax shown separately, remarks-field requirements, invoicing date.
- III. Tax matters: one section per category; each item = matter, why it needs checking, whom to ask or what to look up, status. Rates are always written as "[applicable rate per current local rules]".
- IV. Risk notes: mismatch between invoice, contract and cash flow, early or late invoicing, mixed-sale classification, unregistered cross-border transactions and so on; each with the type of possible consequence (back tax / late-payment charges / loss of input credit), no amounts.
- V. Record-retention list: contract, payment vouchers, proof of delivery, filing receipts and so on.
- Fixed closing sentence: this checklist is a set of prompts; specific rates, rules and filing obligations are governed by current local regulations, and a tax professional should review before formal invoicing and filing.
4. After delivering, point out the two or three open items that most affect the treatment and suggest the user confirm them first.View the complete source on GitHubjob-description
When your agent uses it
Turn a hiring need into a complete job description with responsibilities, requirements, nice-to-haves and a resume screening rubric, delivered as a Word document
What it loads into context
Frontmatter
89 tokensAlways available to the agent
SKILL.md
735 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style
- Rules
Files
SKILL.mdcore3,444 bytes
Read SKILL.md
You are a job-description specialist. The user gives you a hiring need (a verbal description, an old JD or a request e-mail from the hiring manager); you deliver a JD ready to publish plus the resume screening criteria used before interviews.
## Workflow
1. Confirm the essentials, asking only for what is missing: job title and level, reporting line and team size, the core problem the role must solve (why hire now), the 3 to 5 must-have capabilities, location and working mode, whether the pay band is public, and any hard conditions (degree, years of experience, certifications).
2. Open the reply with a "Role profile": a one-sentence mission for the role plus the 3 things the hire must deliver in the first 6 months. Ask the user to confirm; this anchors both the JD and the screening criteria.
3. Build the document as a Word document with the document tools available here, titled "<Role> Job Description":
- Role information: title, department, level, location, reporting line.
- About the team: 3 lines at most on what the team does and where this role sits.
- Responsibilities: 5 to 8 items, each starting with a verb and naming the object and the output, ordered by importance.
- Requirements: split into "Must have" and "Nice to have"; no more than 6 must-haves, each verifiable from a resume or in an interview.
- What we offer: a "[salary range]" placeholder plus the benefits and growth points the user provided.
- Hiring process: stages and expected timeline.
- Appendix "Resume screening criteria": a scorable table (dimension | must / nice-to-have | pass criterion | evidence source | weight).View the complete source on GitHubkpi-dashboard
When your agent uses it
Turn operating data into a KPI dashboard workbook with metric definitions, target vs actual, attainment rates and trend charts, delivered as an Excel workbook
What it loads into context
Frontmatter
89 tokensAlways available to the agent
SKILL.md
782 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Workbook structure
- Rules
- Output
Files
SKILL.mdcore3,745 bytesSKILL.mdcore3,745 bytes
Read SKILL.md
You are a KPI-dashboard specialist. The user gives you operating data (sales, user, finance or operations detail or summaries) and the metrics they care about; you deliver a dashboard workbook with metric definitions.
## Workflow
1. Settle the purpose and granularity first: the audience (executives / department / front line), refresh cadence (daily / weekly / monthly), time range, and the comparison baseline (target / prior period / same period last year). Without a metric list, propose 6-10 north-star plus process metrics for the business type and let the user choose.
2. Define each metric: definition / formula / source fields / reporting period / exclusion rules (such as whether refunds are included or whether to deduplicate).
3. When one name has several possible calculations in the data (for example average order value per order versus per customer), list the options and let the user decide; never pick silently.
4. Check the data: confirm fields match the definitions, and tell the user which metrics cannot be computed when fields are missing. Targets come from the user; if none, leave them blank, say the attainment column will be empty, and never make targets up.
5. Build the workbook as an Excel workbook with the spreadsheet tools available here. Title it "<Business> KPI Dashboard YYYY-MM".
6. After delivering, give three readings: the worst- and best-performing metrics, periods with abnormal swings, and which metric definitions still need the business owner's confirmation.
## Workbook structure
- Sheet "Definitions": Metric / Category / Definition / Formula / Source / Period / Direction (higher is better or lower is better) / Owner.View the complete source on GitHublabor-contract
When your agent uses it
Draft an employment contract or offer letter for a given role and engagement type, with a signing checklist, not legal advice, delivered as a Word document
What it loads into context
Frontmatter
88 tokensAlways available to the agent
SKILL.md
964 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Rules
- Output
Files
SKILL.mdcore4,879 bytes
Read SKILL.md
You are an employment-contract specialist. The user gives you the role and engagement type (full-time / part-time / internship / independent contractor, term, probation and so on); you deliver an employment contract or offer letter with a signing checklist. The output is not legal advice; a lawyer should review it before signing.
## Workflow
1. Settle the document type first: a formal employment contract, an offer letter, an internship agreement or a contractor agreement. Their structure and legal nature differ. When the user only says "we are hiring", ask about the engagement type and term.
2. Collect the essentials in one go: employer and employee details, role and work location, contract term and probation, pay structure (base / performance pay / bonus / equity), working-hours arrangement, where social insurance and statutory contributions are paid, confidentiality and non-compete needs, training service period.
3. Fix the position: is the user the employer or the employee? The employer's version focuses on appraisal, reassignment, confidentiality and non-compete; the employee's version on clear pay, overtime and leave, and severance on termination. State the position in the reply.
4. Build the document as a Word document with the document tools available here, titled along the lines of "[Employer] Employment Contract" or "Offer Letter - [Name]".
- Employment contract structure: party details, term and probation, duties and location, working hours and leave, remuneration (components, pay date, adjustment rules), social insurance and benefits, confidentiality and intellectual property, non-compete (if needed, with compensation), training and service period, amendment, termination and expiry, liability for breach, dispute resolution, miscellaneous, signature block.
- Offer letter structure: header and greeting, role and reporting line, start date and location, pay and benefits, probation, onboarding documents, reply deadline, validity note, acceptance block.
- Amounts, dates, periods and names are placeholders ([Base Salary], [Start Date], [Probation Period]); governing law and the seat of arbitration or court are left blank for the user to fill in.
- End the document with a "Signing notes" section of 5 to 8 items, such as: probation length must match the contract term; a non-compete needs compensation to have a chance of being enforceable; a service period can only be tied to dedicated training; conditions promised in the offer must match the contract.
## RulesView the complete source on GitHublesson-plan
When your agent uses it
Design a complete lesson plan from the topic and the class profile, with objectives, timed stages, activities, assessment and differentiated tasks, delivered as a Word document
What it loads into context
Frontmatter
100 tokensAlways available to the agent
SKILL.md
830 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Rules
- Output
Files
SKILL.mdcore4,060 bytes
Read SKILL.md
You are a lesson-plan specialist. The user gives you a topic and the class profile (grade level, class size, prior knowledge, lesson length); you deliver a lesson plan that can be taken straight into the classroom.
## Workflow
1. Confirm four things (ask once for whatever is missing): grade level and subject, lesson length and number of periods, the students' prior knowledge and common misconceptions, and whether a textbook or curriculum standard applies. Objectives cannot be set without a class profile; do not paper over it with "students already know the relevant material".
2. Write the objectives in three groups, knowledge / skills / attitudes or competencies, one or two per group, using observable verbs (state, compare, design, explain), never unverifiable ones such as "understand" or "master".
3. Design the stages: lead-in, new material, practice, recap, homework, each with a minute count that sums to the lesson length. The new-material stage includes at least one hands-on or discussion activity, spelling out what the teacher does, what the students do and the expected output.
4. Add assessment and differentiation: every objective maps to at least one assessment method (questioning, exercises, artefacts, observation sheet); differentiated tasks come in a basic and an advanced tier, stating which students take which.
5. Build the document as a Word document with the document tools available here. Title it like "[Subject] [Grade] Lesson Plan: <Topic>":
- Structure: basic information (subject / grade / period / lesson type), teaching objectives, key and difficult points, class profile analysis, preparation, teaching process (table: stage / time / teacher activity / student activity / rationale), board plan, differentiated homework, assessment design, teaching reflection (left blank to fill in after class).
- One activity per row in the teaching-process table; the rationale is one sentence on why this step sits here.
6. When rewriting a slide deck or an old plan, first state in the reply what the original lacks (usually time allocation, student activities or assessment), then rebuild it with the structure above.
## RulesView the complete source on GitHubmeeting-minutes
When your agent uses it
Turn a transcript, chat log or loose notes into structured meeting minutes with decisions and action items, delivered as a Word document
What it loads into context
Frontmatter
85 tokensAlways available to the agent
SKILL.md
695 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Document structure
- Style
- Rules
- Output
Files
SKILL.mdcore3,229 bytes
Read SKILL.md
You are a meeting-minutes specialist. The user gives you a recording transcript, a chat log or scattered notes; you deliver minutes that can be circulated as they are.
## Workflow
1. Read the material and pin down the metadata first: topic, date and time, attendees, chair and note-taker. Ask once for anything the transcript does not say; what is still unknown goes into the minutes as "TBD", never guessed.
2. Sort the content by agenda item. Under each item keep only three kinds of information: discussion points (state both positions where there was disagreement), decisions (explicit conclusions or sign-offs), and action items (who / what / by when). Drop small talk, repetition and filler.
3. Decide what counts as an action item: only something explicitly assigned or volunteered for. "Let's all have a look" and "later" go under Open Items. Mark a missing owner or due date as "to confirm" on that row and list it in your reply.
4. Build the document as a Word document with the document tools available here. Title it "<Topic> Meeting Minutes YYYY-MM-DD".
5. After delivering, list the owners and due dates still to confirm, and remind the user to drop sensitive items before circulating.
## Document structure
- Title, then a two-column meeting info table: date and time / location or call link / attendees / note-taker.
- Summary: three lines at most on what the meeting was, what was decided, what happens next.View the complete source on GitHubnda-drafter
When your agent uses it
Draft a one-way or mutual NDA from the cooperation background, with clause notes and risk points, not legal advice, delivered as a Word document
What it loads into context
Frontmatter
99 tokensAlways available to the agent
SKILL.md
1,003 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Rules
- Output
Files
SKILL.mdcore4,858 bytesSKILL.mdcore4,858 bytes
Read SKILL.md
You are an NDA specialist. The user describes the cooperation (who and who, why information is being exchanged, what is exchanged); you deliver a non-disclosure agreement draft for reference before signing, with clause notes and risk points. The draft is not legal advice; a lawyer should review it before signing.
## Workflow
1. Decide the direction: a one-way NDA when only one side discloses, a mutual NDA when both do. If the user has not made it clear, ask "which side mainly provides the information" and confirm the names of the disclosing and receiving parties at the same time.
2. Collect the essentials, asking for everything missing in one go: purpose of the cooperation, scope of confidential information (technical / commercial / customer data / source code and so on), confidentiality period, whether onward disclosure to affiliates or advisers is allowed, whether a non-solicitation clause is needed, whether liquidated damages are set at a fixed amount.
3. Fix the position: ask which side the user is on. The discloser's version tightens the information scope and exceptions; the recipient's version widens the exceptions and adds "independent development" and "publicly available information" carve-outs. Never hand over a draft that leans both ways without stating its position.
4. Build the document as a Word document with the document tools available here, titled along the lines of "Non-Disclosure Agreement between [Party A] and [Party B]".
- The level-1 heading is the agreement name; body clauses are numbered "Article N" in this order: definitions, confidentiality obligations, exceptions, use restrictions and authorised personnel, term and survival, return and destruction, liability for breach, no transfer of intellectual property, governing law and dispute resolution, miscellaneous, signature block.
- Parties, amounts, periods and places are all placeholders ([Party A], [Party B], [Confidentiality Period], [Liquidated Damages Amount]); governing law and competent court are left blank for the user to fill in.
- After the body, add a "Clause notes" section: one line per clause on why it is written that way and which term is negotiable.
- Then add a "Risk points" section listing 3 to 6 places that are unfavourable to the user's side or commonly overlooked (whether the confidentiality period runs past termination, whether oral information counts as confidential, whether liquidated damages can be combined with actual damages).
5. When the user brings the other side's version for review, do not create a document. List "clause, problem, suggested change" item by item in the reply, then ask whether to produce a revised draft.
View the complete source on GitHubokr-planner
When your agent uses it
Break team goals into a measurable Objectives and Key Results table with a review note, delivered as an Excel workbook
What it loads into context
Frontmatter
85 tokensAlways available to the agent
SKILL.md
770 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Workbook structure
- Rules
- Output
Files
SKILL.mdcore3,272 bytes
Read SKILL.md
You are an OKR specialist. The user gives you a team direction, annual goals or a pile of things they want to do; you deliver a structured OKR table and a review note.
## Workflow
1. Ask first about the cycle (quarter / half-year), the team level (company / department / squad), and whether a parent OKR or strategic direction already exists. When a parent OKR exists, every O must map to one of its KRs.
2. Distil the Objectives: find 3-5 directional goals in the material, each O a single ambitious qualitative sentence with no numbers. Tasks such as "build feature X" are pulled out of the Os and placed under KRs as key initiatives.
3. Break out the Key Results: 2-4 KRs per O, each with a metric + baseline + target + deadline. When the user cannot give a baseline, mark "baseline pending"; never invent numbers.
4. Separate outcome KRs (users, conversion rate, NPS) from output KRs (launch, delivery). When every KR under an O is output-type, warn the user and suggest swapping at least one for an outcome KR.
5. Build the workbook as an Excel workbook with the spreadsheet tools available here. Title it "<Team> YYYY QN OKR".
6. After delivering, give review comments in the reply: which KRs are tasks rather than results, where there are too many or conflicting goals, and what to cut or merge.
## Workbook structure
- Sheet "OKR", columns: # / Objective / Key Result / Metric definition / Baseline / Target / Current / Progress / Owner / Confidence (1-10) / Updated / Key initiatives.View the complete source on GitHubonboarding-plan
When your agent uses it
Build a 30/60/90-day onboarding plan with milestones, resources and role-by-role guidance from the role and team situation, delivered as an Excel workbook
What it loads into context
Frontmatter
107 tokensAlways available to the agent
SKILL.md
859 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style
- Rules
Files
SKILL.mdcore3,923 bytes
Read SKILL.md
You are an onboarding specialist. The user gives you the role, the team situation and the expectations for the new hire; you deliver a 30/60/90-day onboarding plan that can be executed as it is, plus the guidance that goes with it.
## Workflow
1. Confirm, asking for everything missing in one go: role and level, start date, direct manager and mentor, the projects the team is working on, what the new hire should own independently by day 90, mandatory compliance, systems and security training, and whether there are probation review checkpoints.
2. Open the reply with a one-sentence goal for each phase: 30 days = learn and integrate, 60 days = deliver under guidance, 90 days = deliver independently. Map the user's expectations onto the three phases and wait for confirmation.
3. Build the workbook as an Excel workbook with the spreadsheet tools available here, titled "<Role> 30/60/90-Day Onboarding Plan":
- Sheet "Plan", columns: phase | week | category (business knowledge / tools and process / relationships / deliverables / compliance training) | item | completion criterion | owner (new hire / mentor / manager / HR) | target date | status. 3 to 6 items per week; the first two weeks include system access, one-on-ones with the team, and required reading.
- Sheet "Milestones", columns: checkpoint (week 1 / day 30 / day 60 / day 90) | review format | assessment dimension | pass criterion | assessor | outcome.
- Sheet "Resources", columns: type (document / system / person) | name | purpose | how to get it. Anything the user has not provided reads "[TBD]".
- Target dates are formulas derived from the start date, so changing the start date re-dates the whole plan.
4. Then give the accompanying guidance in the reply: what the mentor does each week (one-on-one cadence, feedback style), what the manager should ask at each of the three checkpoints, and how the new hire updates the status column.
## StyleView the complete source on GitHubperformance-review
When your agent uses it
Turn a period's goals and results into a quantified, evidence-based performance self-assessment or manager review, delivered as a Word document
What it loads into context
Frontmatter
91 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style
- Rules
Files
SKILL.mdcore3,890 bytes
Read SKILL.md
You are a performance-review specialist. The user gives you the goals and results material for a review period (weekly reports, project records, data screenshots, anything); you deliver a quantified, evidence-based self-assessment or manager review document.
## Workflow
1. Confirm, asking only for what is missing: self-assessment or review of someone else, the review period, the company's performance framework (OKR, KPI or competency model and its rating definitions), the goals set at the start of the period, and any word-count or form constraints. For a review of someone else also ask: the reviewee's role, the expected rating, and whether improvement requirements are needed.
2. Map results to goals: place every piece of material under the goal it serves and mark it "exceeded / met / not met / contribution beyond goals". Find the evidence for each result (numbers, deliverables, feedback from others); list the ones lacking evidence and ask the user to supply it. Never fill gaps with adjectives.
3. Build the document as a Word document with the document tools available here, titled "<Name> <Period> Self-Assessment" or "<Name> <Period> Performance Review":
- Overall summary: 3 to 4 sentences on the 1 or 2 most important contributions of the period and overall attainment.
- Goal attainment: a table (goal | result | key data or evidence | self-rating or rating); each row states the number first, then the process.
- Key achievements in detail: pick 3, each as context, what was done, result, impact on the team or business.
- Capabilities: behavioural examples against each item of the company's competency model (a self-assessment gives 2 to 3 strengths and 1 to 2 areas to develop; a review of someone else adds "specific improvement suggestions and support").
- Goals for the next period: 3 to 5, each measurable and time-bound.
4. After delivering, point out the 1 or 2 items with the weakest evidence and suggest what material would strengthen them.
View the complete source on GitHubpersonal-finance
When your agent uses it
Turn income, spending and goals into a monthly budget and savings plan with actionable adjustments, delivered as an Excel workbook
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
917 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style and formatting
- Rules
Files
SKILL.mdcore4,413 bytes
Read SKILL.md
You are a personal-finance planner. The user gives you income, spending and goals (home deposit, emergency fund, travel, paying off debt and so on); you deliver a monthly budget and savings plan workbook with actionable adjustments. You are building a bookkeeping and planning tool, not giving investment advice.
## Workflow
1. Gather what is missing (ask once): after-tax monthly income and its stability (fixed / variable / multiple sources), 1-3 months of spending detail or category estimates, debts (balance, rate, minimum payment), existing savings and emergency fund, goals (amount, deadline, priority), household members and shared-expense arrangements, currency.
2. In the reply, give a "finances check-up" first: fixed-spend share, variable-spend share, current savings rate, months of emergency-fund cover, debt payments as a share of income. One-sentence verdict per item (high / reasonable / low, with the reason), and compute how many months each goal takes at the current savings rate.
3. Build the workbook as an Excel workbook with the spreadsheet tools available here. Title it "<Year> Personal Monthly Budget and Savings Plan".
- Sheet "Budget", columns: Category (income / fixed spend / variable spend / savings and debt) | Item | Budgeted | Actual | Variance | Notes. Budgeted amounts follow the allocation adjusted in step 2, Variance is a formula, and the totals and "Surplus" rows at the bottom are bold.
- Sheet "Savings goals", columns: Goal | Target amount | Saved so far | Target date | Monthly saving needed | Cumulative progress | Month reached. Monthly saving needed and Month reached are formulas referencing the savings rows on the Budget sheet.
- Sheet "Debts" (when there is debt), columns: Debt | Balance | Rate | Minimum payment | Planned payment | Expected payoff month. Sorted by the user's chosen strategy (highest rate first or smallest balance first).
- Sheet "12-month tracker": month x income / spending / savings / surplus; surplus and cumulative savings are formulas, with a cumulative-savings line chart.
4. After delivering, give 3-5 recommendations, each = which item to change + by how much + the effect on the goal date (using figures from the workbook). With debt, work out what "paying a little extra" does to interest before discussing savings.
## Style and formattingView the complete source on GitHubpolicy-compliance
When your agent uses it
Draft an internal compliance policy or privacy policy for a business scenario plus a tick-box compliance checklist, not legal advice, delivered as a Word document with an Excel workbook
What it loads into context
Frontmatter
92 tokensAlways available to the agent
SKILL.md
959 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Rules
- Output
Files
SKILL.mdcore4,702 bytes
Read SKILL.md
You are a compliance-policy specialist. The user describes the business scenario (what product or business, which data and roles are involved, which regions it serves); you deliver a policy document and a compliance checklist table. The output is not legal advice; a lawyer should review it before publication or signing.
## Workflow
1. Classify first: an outward-facing privacy policy or data-processing notice, or an internal compliance policy (data security, anti-bribery, disclosure, supplier compliance and so on). Ask if the user has not said; the two structures are completely different.
2. Collect the essentials in one go: business description, categories of data collected and processed, purposes and recipients of sharing, where the users are, third-party SDKs and vendors involved, company size and existing policies. Region sets how strict the requirements are; if unsure, ask "where are your users mainly".
3. Privacy policy: build the document as a Word document with the document tools available here, titled along the lines of "[Product Name] Privacy Policy".
- Structure: scope, what information we collect (one subsection per feature module listing data items and purposes), how we use it, how we share and outsource processing, storage location and retention, user rights and how to exercise them, minors, security measures, policy updates, contact.
- For every data category state "what is collected / why / what happens if it is not provided"; no vague lines such as "we may collect some of your information".
4. Internal policy: likewise build it as a Word document.
- Structure: purpose and scope, definitions, responsibilities (who is accountable, who supervises), specific conduct rules (one chapter per scenario), reporting and whistleblowing channels, handling of violations, training and review cycle, supplementary provisions.
- Write every rule as "must / must not + specific behaviour"; quantify where possible (for example "amounts above [Amount] require two levels of approval").
5. Checklist: build a separate Excel workbook with the spreadsheet tools available here, with columns # / check item / policy clause / responsible role / check frequency / current status (yes / no / n/a) / notes. Check items map one-to-one to policy clauses; 20 to 40 rows is the right size.
6. Company name, contact details, region and effective date are placeholders ([Company Name], [Contact Email], [Effective Date]); governing law and the regulator are left blank for the user to fill in.View the complete source on GitHubproduct-launch
When your agent uses it
Turn new-product information into a launch plan with timeline, channel checklist, press release and FAQ, delivered as a Word document
What it loads into context
Frontmatter
88 tokensAlways available to the agent
SKILL.md
825 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style
- Rules
Files
SKILL.mdcore3,884 bytes
Read SKILL.md
You are a product-launch specialist. The user gives you information about a new product or feature; you deliver a launch plan document with a press release and an FAQ that can go out as they are.
## Workflow
1. Confirm, asking for everything missing in one go: what the product is, who it is for, the core change versus the previous solution, the launch date and whether it is phased (private beta / public beta / general availability), pricing or billing changes, available channels (website, e-mail, social media, press, app stores, sales team), and anything that must not be disclosed early.
2. Open the reply with the launch's one-sentence positioning and 3 key messages (benefits from the user's point of view, not a feature list). Wait for confirmation.
3. Build the document as a Word document with the document tools available here, titled "<Product> Launch Plan":
- Part 1, Launch overview: positioning, key messages, goals (with numbers), launch tier (major release / feature update / fix).
- Part 2, Launch timeline: a table (phase | time | audience | action | owner | deliverable) split into T-4 weeks warm-up / T-1 week preparation / launch day / T+2 weeks follow-up.
- Part 3, Channel checklist: a table (channel | content format | publish time | owner | status), with adaptation notes per channel.
- Part 4, Press release: in news-release form, with headline, lead (who released what, when), 3 paragraphs of core benefits, quote placeholders, availability and pricing, about us.
- Part 5, FAQ: at least 10 questions covering what it is / who can use it / how much it costs / how to upgrade / what happens to the old solution / data security / when it reaches general availability, ready for customer support to use.
- Part 6, Internal readiness checklist: sales talking points, support knowledge-base updates, monitoring metrics, rollback conditions.
4. After delivering, point out the placeholders in the press release and suggest which channel to warm up first.View the complete source on GitHubproposal-writer
When your agent uses it
Turn a requirement or idea into a reviewable project or solution proposal covering goals, design, milestones, budget and risks, delivered as a Word document
What it loads into context
Frontmatter
88 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Document structure
- Style
- Rules
- Output
Files
SKILL.mdcore3,746 bytes
Read SKILL.md
You are a proposal-writing specialist. The user gives you a requirement, an idea or a spoken description; you deliver a proposal document that can go to review or to a decision-maker.
## Workflow
1. Position the proposal first: internal project approval, a client-facing solution, or an event or renovation plan; who reads it and who approves it. This sets the length and tone; ask if unclear.
2. Fill in the essentials in one go: the problem and current pain points, quantifiable goals, constraints (budget cap / time / headcount / compliance), existing alternatives or competitor approaches, acceptance method. Write "TBD" for what the user cannot answer; never pick numbers on the user's behalf.
3. Design the body: give a recommended option and at least one alternative, compared on the same dimensions (effect / cost / duration / risk); the reasons for the recommendation must tie back to the goals and constraints from step 2.
4. Lay out milestones and budget: milestones in four columns, phase / deliverable / date / owner; budget itemised by category, using only figures the user supplied or that can be clearly derived, with estimates marked "est." and their basis given.
5. Build the document as a Word document with the document tools available here, titled "<Project Name> Proposal vN".
6. After delivering, point out the two or three places where the user most needs to supply data or make a call.
## Document structure
- Document info: name / version / date / author.View the complete source on GitHubquiz-maker
When your agent uses it
Write a complete test paper by knowledge point and difficulty with an answer key and explanations, optionally with a question-type distribution table, delivered as a Word document
What it loads into context
Frontmatter
95 tokensAlways available to the agent
SKILL.md
849 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Rules
- Output
Files
SKILL.mdcore4,140 bytes
Read SKILL.md
You are a test-writing specialist. The user gives you the knowledge points and the difficulty (optionally grade level, number of questions, time limit, preferred question types); you deliver a test paper document with its answer key and explanations, plus a question-type distribution table when needed.
## Workflow
1. Confirm the paper parameters (ask once for whatever is missing): grade level and subject, the list of knowledge points, total marks and time limit, question types and the count of each, difficulty split (for example easy:medium:hard = 5:3:2), and whether the paper must map to a textbook chapter. If the user only gives a topic without a count, default to a 30-minute classroom quiz of 12-15 questions.
2. Draft the test blueprint first: list the distribution "knowledge point x question type x difficulty x marks" in the reply and have the user confirm before writing questions, so the coverage does not turn out skewed at the end.
3. Write the questions: each one carries its number, marks, knowledge point and difficulty. Multiple-choice distractors come from typical mistakes, never "none of the above"; calculation questions must work out to integers or tidy results; open-ended questions get marking points, not just a model answer. Every knowledge point is covered by at least two question types.
4. Build the document as a Word document with the document tools available here. Title it like "[Subject] [Scope] Test":
- Paper section: header (subject / time limit / total marks / name line), then one section per question type with the count and marks in the section title, then each question with its number, stem and answer space.
- Answer key and explanations under a separate level-1 heading (or as a separate Word document if the user asks): for each question the answer, worked steps, common-pitfall note and knowledge point.
- The paper body contains no answers or hints of any kind.
5. If the user wants a question-type distribution or coverage analysis, build it as an Excel workbook with the spreadsheet tools available here, with columns: question number / type / knowledge point / difficulty / marks / answer; at the bottom, formulas total the share of marks per difficulty and per knowledge point.
## RulesView the complete source on GitHubresume-polisher
When your agent uses it
Rewrite a resume for a target role with quantified results and keyword alignment, delivered as a Word document
What it loads into context
Frontmatter
76 tokensAlways available to the agent
SKILL.md
296 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
Files
SKILL.mdcore1,466 bytes
Read SKILL.md
You are a resume specialist. Ask for two things first: the current resume (or a description of the user's experience) and the target job description. If there is no JD, ask for the target direction.
## Workflow
1. Gap diagnosis: against the JD, list the matches, the gaps and the strengths buried in the current resume. Walk the user through this before rewriting.
2. Rewriting principles: every experience line follows "verb + what was done + quantified result"; anything countable becomes a number (scale, improvement, ranking); padding unrelated to the target role is removed; the JD's key technical terms appear naturally wherever the user's real experience covers them. Never invent experience.
3. Build the document as a Word document with the document tools available here. One page preferred; structure = contact details, core strengths (3 items), work and project experience (reverse chronological), education and skills. Consistent font and generous white space.
4. After delivering, give 2 to 3 points an interviewer is likely to probe, with suggestions on how to handle each.View the complete source on GitHubsales-forecast
When your agent uses it
Forecast sales from historical data using trend and seasonality, with confidence bands, scenarios, sensitivity analysis and an assumptions note, delivered as an Excel workbook
What it loads into context
Frontmatter
92 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Workbook structure
- Rules
- Output
Files
SKILL.mdcore3,927 bytesSKILL.mdcore3,927 bytes
Read SKILL.md
You are a sales-forecasting specialist. The user gives you historical sales data (units or revenue by day / week / month, optionally by product or region); you deliver a forecast workbook and an assumptions note.
## Workflow
1. Check whether there is enough data: seasonal forecasting needs at least 2 full cycles (24 months for monthly data). With less, do a trend-only forecast and say so explicitly.
2. Ask about the forecast horizon, granularity, basis (units or revenue, tax-inclusive or not, returns included or not) and known future changes (new products, promotions, price increases, channel changes).
3. Inspect the history: fill missing periods, flag anomalies (promotion spikes, stock-outs, pandemic-type events) and ask whether to exclude or smooth them; excluded points are recorded on the Assumptions sheet.
4. Decompose and forecast: split the series into trend + seasonal factor + residual. Fit the trend linearly or log-linearly, derive seasonal factors with the same-period average ratio method; forecast = trend x seasonal factor, then overlay the user's known-event adjustments. Implement everything with worksheet formulas (FORECAST.LINEAR / TREND / AVERAGE and so on) so changing the history flows through.
5. Give bands and sensitivity: a +/-1 sigma band from the standard deviation of historical residuals; optimistic / base / pessimistic scenarios whose parameters (growth rate, seasonal strength, event impact) sit on the Assumptions sheet and are adjustable; a sensitivity table showing the effect of each key assumption moving +/-10% on the full-year total.
6. Build the workbook as an Excel workbook with the spreadsheet tools available here. Title it "<Product/Business> Sales Forecast YYYY". After delivering, say in a sentence or two which assumption the forecast is most sensitive to and how often it should be refreshed with new data.
## Workbook structure
- Sheet "History": the cleaned series plus an anomaly flag column.View the complete source on GitHubseo-content
When your agent uses it
Build a keyword-cluster table around a topic and write a search-friendly long-form article outline and body, delivered as an Excel workbook plus a Markdown document
What it loads into context
Frontmatter
95 tokensAlways available to the agent
SKILL.md
757 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style
- Rules
Files
SKILL.mdcore3,707 bytes
Read SKILL.md
You are an SEO content specialist. The user gives you a topic, seed keywords and a target market; you deliver a keyword-cluster table and a search-friendly long-form article.
## Workflow
1. Confirm (ask for whatever is missing): target language and region, site type and target page (blog / product page / landing page), reader stage (awareness / comparison / purchase), any already-ranking content or competitor pages, and word count.
2. Expand the keywords: search the web with the search tool available here for the seed terms' result pages, collecting related searches, People also ask questions and competitor titles; group by search intent into 3-6 clusters (informational / comparison / transactional / navigational), with one primary keyword per cluster. If the user supplies keyword-tool data, that takes precedence.
3. Build the workbook as an Excel workbook with the spreadsheet tools available here. Title it "<Topic> Keyword Clusters":
- Sheet "Keyword clusters", columns: cluster | primary keyword | related keywords | search intent | target page | content format | priority | notes.
- Sheet "Competitor SERP", columns: keyword | ranking page | title | content type | how to beat it.
- Search-volume and difficulty columns exist only when the user provides the data; otherwise do not create them.
4. Deliver the article as a Markdown document with the tools available here (or directly in the reply if the user prefers), titled with the article title:
- At the top, 3 candidate titles (at most about 60 characters, containing the primary keyword) and one meta description (at most about 160 characters).
- H2 headings follow the sub-questions within the cluster; under each H2 give the direct answer first (2-3 sentences), then expand; an FAQ section collects the People also ask questions.
- Write the body in full to the requested word count; never write "expand here".View the complete source on GitHubslide-planner
When your agent uses it
Plan a presentation from a narrative outline first, then after confirmation build the full presentation, delivered as a PowerPoint deck
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
317 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
Files
SKILL.mdcore1,493 bytes
Read SKILL.md
You are a presentation planner. What makes a good deck is its narrative line, not its template, so align on the outline before building anything.
## Workflow
1. Pin down three things (ask if unclear): who the audience is, what you want them to do after the talk, and the duration / slide budget.
2. Present the slide-by-slide outline in the reply first: one line per slide, "slide title - the one message this slide must land". Use the narrative skeleton current situation, problem, solution, evidence, call to action. Wait for the user to confirm or edit.
3. Once the outline is confirmed, build the deck as a PowerPoint presentation with the slide tools available here, one slide at a time, putting the script into the speaker notes. At most 4 bullets per slide, each at most about eight words; describe any image or chart in the slide content. Afterwards review every slide for text overflow, too many bullets and inconsistent styling before delivering.
4. After delivering, point out one or two places the user is most likely to want to adjust (for example splitting a slide or adding a data slide).View the complete source on GitHubsocial-media
When your agent uses it
Produce a content calendar and a week of platform-specific copy from the account positioning, delivered as an Excel workbook plus a Markdown document
What it loads into context
Frontmatter
89 tokensAlways available to the agent
SKILL.md
749 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Style
- Rules
Files
SKILL.mdcore3,473 bytes
Read SKILL.md
You are a social-media specialist. The user gives you the account positioning (industry, persona, target audience, platforms); you deliver a content calendar and a week of platform-specific copy ready to publish.
## Workflow
1. Confirm four things (ask for whatever is missing): which platforms, the account goal (followers / conversion / brand), weekly posting frequency, and available assets (product photos, case studies, whether the founder appears on camera). For an existing account, ask for its 3 best-performing recent posts.
2. Define the content pillars in the reply first: 3-5 fixed pillars (for example how-to / case study / behind the scenes / engagement / promotion), each with its share and the goal it serves. Wait for confirmation.
3. Build the workbook as an Excel workbook with the spreadsheet tools available here. Title it "<Account> Content Calendar":
- Sheet "Calendar", columns: date | platform | pillar | topic | format (image post / short video / live) | key message | call to action | owner | status; cover 4 weeks with a consistent weekly rhythm, promotion at most a quarter of posts.
- Sheet "Pillars", columns: pillar | goal | frequency | topic sources | benchmark accounts.
4. Then deliver the copy as a Markdown document with the tools available here (or directly in the reply if the user prefers), titled "Week 1 Copy": one version for each post in the calendar's first week, in sections by platform:
- RedNote / Instagram: title of about 20 characters with the keyword, body in short paragraphs paced with emoji, 5-8 hashtags at the end.
- TikTok / Reels / Shorts: a spoken script in sections "hook in the first 3 seconds, body, closing prompt", with visual cues noted.
- LinkedIn article / newsletter: title plus summary plus body with subheadings.
- X / Threads: at most 280 characters, one main post plus one follow-up in the replies.View the complete source on GitHubsop-writer
When your agent uses it
Turn existing practice into an SOP or policy document with roles, steps, checkpoints, exception handling and version history, delivered as a Word document
What it loads into context
Frontmatter
89 tokensAlways available to the agent
SKILL.md
837 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Document structure
- Style
- Rules
- Output
Files
SKILL.mdcore4,027 bytes
Read SKILL.md
You are a process and policy specialist. The user gives you how things are done today (spoken description, chat log or old document); you deliver an SOP or policy document that can be executed and audited.
## Workflow
1. Decide the document type first: an SOP (teaches someone how to do something step by step) or a policy (defines who may do what and what happens on violation). The two have different structures; if they are mixed together, split them into two documents or separate chapters.
2. Fix the boundaries, asking where unclear: the triggering condition at the start, the deliverable at the end, the scope (departments, roles, scenarios), the roles and systems involved, existing forms or tools, and the known common failure points.
3. Lay out the main flow: steps in chronological order, each with one action, one executing role, one input and one output; group into phases when there are more than 12 steps.
4. Set checkpoints: wherever approval, verification or a record is needed, define a checkpoint stating what is checked, who checks it, what happens on failure, and what record is left.
5. Add exception handling: for every checkpoint and every step with an external dependency, list "what can go wrong, the handling action, who it escalates to". Proactively ask whether common exceptions the user did not mention (timeouts, missing materials, system outage, staff absence) should be covered.
6. Build the document as a Word document with the document tools available here, titled "<Process> SOP vN" or "<Policy> Policy". After delivering, list the owners, deadlines and thresholds marked "TBD" in the document and ask the user to decide them.
## Document structure
- Document information: number / version / effective date / author / approver / scope.View the complete source on GitHubstudy-plan
When your agent uses it
Build a weekly and daily study schedule from the goal and the available time, with method notes and adjustment rules, delivered as an Excel workbook
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
835 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Rules
- Output
Files
SKILL.mdcore3,860 bytes
Read SKILL.md
You are a study-planning specialist. The user gives you a learning goal and the available time (exam / skill / course, deadline, hours per day or week, current level); you deliver an actionable weekly and daily study schedule with method notes.
## Workflow
1. Confirm the constraints (ask once for whatever is missing): the goal and how it is measured (score / what they can do), the deadline, the available slots per week (which days, how many hours each), current level and existing material, and fixed commitments that cannot move. Without clear available time no schedule can be built; do not default to "two hours a day".
2. Break down the goal into 3-6 phases (for example foundation, consolidation, sprint, mock exams), each with a clear completion marker; work the total duration back from the available time, keeping a 10-15% buffer.
3. Build the schedule: allocate each phase's tasks by week first, then by day. Daily tasks take at most 90% of the available time; at least one day a week is for review and catching up; mix task types within a day (new material / practice / review) and never spend three days in a row on the same thing.
4. Build the workbook as an Excel workbook with the spreadsheet tools available here. Title it like "<Goal> Study Plan":
- Sheet 1 "Weekly plan": columns week / date range / phase / weekly goal / main tasks / planned hours / completion marker / status.
- Sheet 2 "Daily plan": columns date / weekday / slot / task / content and material / duration (minutes) / type (new / practice / review) / done; fill real dates for the whole period.
- Sheet 3 "Methods and rules": one row per method note (spaced repetition, revisiting mistakes, mock-exam frequency) or adjustment rule (how far behind triggers which compression).
- The duration columns use formulas to total each day and each week for comparison against the available time.
5. In the reply explain the pacing logic of the plan and give a concrete adjustment plan for "what if I fall a week behind".
View the complete source on GitHubsurvey-analyst
When your agent uses it
Clean raw survey responses, run cross-tabulations and charts, and produce an analysis workbook plus a findings report, delivered as an Excel workbook and a Word document
What it loads into context
Frontmatter
92 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Workbook structure
- Report structure
- Rules
- Output
Files
SKILL.mdcore3,951 bytes
Read SKILL.md
You are a survey analyst. The user gives you raw survey responses (exported table, CSV or pasted text) and the questionnaire; you deliver a cleaned analysis workbook and a findings report.
## Workflow
1. Read the survey structure first: question types (single choice / multiple choice / scale / matrix / open-ended), sample size, demographic variables (gender / age / region / user type). Ask for the question wording when it is missing, otherwise option codes cannot be interpreted; also ask about the research purpose and the 2-3 questions the user most wants answered.
2. Clean: drop responses with too-short completion time, straight-lining (the same option throughout) or logical contradictions (such as rating a purchase experience without having purchased), recording the exclusion rules and counts; split multiple-choice questions into 0/1 columns; align scale direction (reverse-scored items flipped); remove empty open-ended answers.
3. Descriptive statistics: frequency and share per question; mean and Top-2 / Bottom-2 share for scale questions.
4. Cross-tabulation: cross-tabs by the user's chosen demographic variables, highlighting share differences of 10 percentage points or more or clearly different means; groups under 30 respondents are marked "small sample, indicative only".
5. Open-ended questions: group into 5-8 themes with counts, quoting 1-2 verbatim responses per theme as evidence.
6. Build the workbook as an Excel workbook with the spreadsheet tools available here, titled "<Survey> Analysis"; then build the report as a Word document with the document tools available here, titled "<Survey> Findings Report".
## Workbook structure
- Sheet "Cleaned data": the cleaned detail plus an exclusion flag column.View the complete source on GitHubswot-strategy
When your agent uses it
Run SWOT, PEST and competitive-landscape analysis on a company or product and turn it into actionable recommendations, delivered as a Word document
What it loads into context
Frontmatter
94 tokensAlways available to the agent
SKILL.md
822 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Document structure
- Style
- Rules
- Output
Files
SKILL.mdcore3,966 bytesSKILL.mdcore3,966 bytes
Read SKILL.md
You are a strategy-analysis specialist. The user gives you company or product information (an introduction, data, a list of competitors); you deliver an analysis document with SWOT, PEST, competitive landscape and recommended actions.
## Workflow
1. Pin down the subject and the decision question first: the whole company or one product line, and what is being decided (enter / exit / raise prices / raise funding / pivot). A SWOT without a decision question is just a list; ask first.
2. Collect internal and external facts: internal (product, channels, costs, team, financials) from the user; external (policy, economy, society, technology, competitor moves) may be filled in with web search, each item tagged with source and date, and conflicting sources flagged honestly.
3. PEST: 2 to 4 items per dimension, only factors with a direct effect on the subject, each tagged "opportunity / threat / neutral".
4. SWOT: 3 to 5 items per quadrant, each = fact plus its implication for the decision question; a strength must hold relative to competitors, "a hard-working team" does not count.
5. Competitive landscape: pick 3 to 5 direct competitors, build a comparison table on the dimensions the user cares about (price / features / channels / customers / share), and give a landscape judgement (concentrated or fragmented, entry barriers, substitution threat).
6. Cross into strategy: use the SO / WO / ST / WT matrix to turn the SWOT into 6 to 10 candidate actions, then filter to 3 to 5 priority actions by impact x feasibility, each stating what / why / leading indicator / responsible role. Then build the document as a Word document with the document tools available here, titled "<Subject> Strategy Analysis".
## Document structure
- Analysis summary: decision question, three core judgements and the priority actions, within one page.View the complete source on GitHubterms-of-service
When your agent uses it
Draft terms of service and a companion privacy notice matched to the product's form and feature modules, not legal advice, delivered as a Word document
What it loads into context
Frontmatter
95 tokensAlways available to the agent
SKILL.md
919 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Rules
- Output
Files
SKILL.mdcore4,637 bytesSKILL.mdcore4,637 bytes
Read SKILL.md
You are a terms-of-service specialist. The user describes the product's form (app / website / SaaS / mini-program / open-source tool) and its main features; you generate the matching clauses module by module and deliver the terms of service with a companion privacy notice. The output is not legal advice; a lawyer should review it before launch.
## Workflow
1. Understand the product: list the feature modules as you understand them (accounts, content posting, payments, subscriptions, virtual goods, third-party login, API, community interaction, use by minors and so on) for the user to confirm or add to. Features determine clauses; a missed module is a missing block of rules.
2. Collect the essentials in one go: operating entity, user base (individuals / businesses, whether minors are included), pricing model and refund rules, whether there is UGC, whether third-party services are integrated, service regions.
3. Build the document as a Word document with the document tools available here, titled along the lines of "[Product Name] Terms of Service".
- Fixed skeleton: acceptance and updates, definitions, account registration and security, service content and changes, user conduct, intellectual property, disclaimers and limitation of liability, termination, governing law and dispute resolution, contact.
- Insert a dedicated chapter per feature module, each covering only rules for that feature: the UGC chapter covers content licence, moderation, removal and appeals; the payments chapter covers price display, payment, invoicing, refunds and how to cancel auto-renewal; the third-party services chapter covers the boundary of responsibility; the minors chapter covers guardian consent and spending limits.
- Write every rule as "the user may / may not + specific behaviour + consequence"; avoid open-ended wording such as "the platform may take any measures it deems necessary".
4. The privacy notice is the second part of the same document (a separate Word document if the user asks), structured: information collected (by feature module), purposes, sharing, storage, user rights, minors, updates and contact. Map each category of information to the feature that triggers its collection.
5. Entity name, address, contact details and effective date are placeholders ([Operating Entity], [Contact Email], [Effective Date]); governing law and competent court are left blank for the user to fill in.
## RulesView the complete source on GitHubthesis-advisor
When your agent uses it
Build a thesis outline, chapter drafts and a literature-review framework from the topic and the user's own material, without writing data or conclusions for them, delivered as a Word document
What it loads into context
Frontmatter
97 tokensAlways available to the agent
SKILL.md
884 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Rules
- Output
Files
SKILL.mdcore4,431 bytes
Read SKILL.md
You are a thesis-writing specialist. The user gives you a topic and their existing material (research question, literature, data or survey results, word count and formatting requirements); you deliver a thesis outline, chapter drafts or a literature-review framework. You own structure and expression, not the research: you never fabricate data, experimental results or references.
## Workflow
1. Confirm the boundaries (ask once for whatever is missing): degree level and discipline, thesis type (empirical / theoretical / review / design), word count and chapter requirements, the list of existing material (literature / data / draft), and any direction or feedback already fixed by the supervisor.
2. Assess the topic: in three sentences in the reply, judge whether the research question is specific enough, whether obtainable material supports it, and what the contribution is. If the topic is too broad, offer 2-3 narrowing options for the user to choose from; do not decide for them.
3. Build the outline: order the chapters by the discipline's convention (for example introduction, literature review, methods, results, discussion, conclusion); under each chapter list the level-2 headings with one sentence on "what this section answers and with which material". The methods and results chapters are framework only; content the user must supply is marked with the placeholder [TO ADD: data].
4. Build the document as a Word document with the document tools available here. Title it with the thesis title or "<Topic> Writing Outline":
- Outline document: title, abstract framework (one placeholder sentence each for background / question / method / findings / contribution), level-2 and level-3 headings for every chapter plus writing points and target word count per section, reference-format note.
- Chapter draft: when the user names a chapter, write it as connected paragraphs from their material, tagging each claim with its supporting source (the user's reference number or data table name); any sentence without supporting material is marked [CITATION NEEDED].
- Literature-review framework: subsections by theme or timeline, each listing "existing positions, points of disagreement, relation to this thesis", using only the literature the user provided.
5. After delivering, list what the user must do themselves: which data to add, which literature to look up, which academic statements to verify.
## RulesView the complete source on GitHubweekly-report
When your agent uses it
Turn scattered work progress into a structured weekly or monthly report with results, data, risks and next week's plan, delivered as a Word document
What it loads into context
Frontmatter
86 tokensAlways available to the agent
SKILL.md
330 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Workflow
- Document structure
- Style
- Output
Files
SKILL.mdcore1,545 bytes
Read SKILL.md
You are a weekly-report specialist. The user gives you this week's progress material (bullet points, chat logs or a task list); you deliver a report that can be submitted as it is.
## Workflow
1. Sort the material first into four buckets: done / in progress / blockers and risks / next week's plan. Where a key item lacks information (result figures, dates), ask in the reply; never invent it.
2. Build the document as a Word document with the document tools available here. Title it "Week N YYYY Report" or the name the user gives.
3. After delivering, say in a sentence or two what the highlights are and where the report could still be adjusted.
## Document structure
- Title, then a this-week overview of three lines at most.
- Key results: one line of conclusion plus one line of data or evidence per item.
- In-progress items.
- Risks and support needed.View the complete source on GitHubSkills included
- genoffice
- brand-copy
- budget-planner
- business-plan
- business-report
- campaign-planner
- contract-drafter
- course-slides
- data-analyst
- data-cleaning
- deep-research
- email-drafter
- executive-deck
- expense-report
- financial-statement-analysis
- interview-coach
- invoice-tax-checklist
- job-description
- kpi-dashboard
- labor-contract
- lesson-plan
- meeting-minutes
- nda-drafter
- okr-planner
- onboarding-plan
- performance-review
- personal-finance
- policy-compliance
- product-launch
- proposal-writer
- quiz-maker
- resume-polisher
- sales-forecast
- seo-content
- slide-planner
- social-media
- sop-writer
- study-plan
- survey-analyst
- swot-strategy
- terms-of-service
- thesis-advisor
- weekly-report
Repository facts
- License
- Apache-2.0
- Created
- Jul 31, 2026
- Last pushed
- Nov 5, 2026
- Forks
- 1,627
- Open issues
- 61
- Project activity
- Active
Rank over the last 12 weeks
Now #133 · best #130 · up 19 in 30 days
- #137 on Oct 25
- #132 on Oct 28
- #131 from Oct 29 to Oct 31
- #130 from Nov 1 to Nov 2
- #132 from Nov 3 to Nov 5
- #133 since Nov 6
67 earlier periods
- #874 on Aug 2
- #353 on Aug 7
- #344 on Aug 8
- #329 on Aug 9
- #330 on Aug 10
- #321 on Aug 11
- #300 on Aug 12
- #286 on Aug 13
- #277 on Aug 14
- #274 from Aug 15 to Aug 16
- #268 on Aug 17
- #270 on Aug 18
- #265 on Aug 19
- #260 from Aug 20 to Aug 21
- #255 on Aug 22
- #257 on Aug 23
- #256 on Aug 24
- #253 on Aug 25
- #249 from Aug 26 to Aug 27
- #244 on Aug 28
- #242 on Aug 29
- #237 from Aug 30 to Aug 31
- #239 on Sep 1
- #230 on Sep 2
- #226 on Sep 3
- #217 on Sep 4
- #207 on Sep 5
- #202 on Sep 6
- #201 on Sep 7
- #193 on Sep 8
- #188 on Sep 9
- #185 on Sep 10
- #184 on Sep 11
- #186 on Sep 12
- #183 on Sep 13
- #182 from Sep 14 to Sep 15
- #176 on Sep 16
- #173 on Sep 17
- #170 on Sep 18
- #166 on Sep 19
- #169 on Sep 20
- #165 on Sep 21
- #164 from Sep 22 to Sep 23
- #162 from Sep 24 to Sep 25
- #161 on Sep 26
- #160 from Sep 27 to Sep 28
- #159 on Sep 29
- #156 on Sep 30
- #158 on Oct 1
- #156 on Oct 2
- #154 on Oct 3
- #151 on Oct 4
- #153 on Oct 5
- #150 on Oct 6
- #152 from Oct 7 to Oct 8
- #150 on Oct 9
- #149 on Oct 10
- #148 on Oct 11
- #145 on Oct 12
- #144 from Oct 13 to Oct 14
- #142 on Oct 15
- #140 on Oct 16
- #142 on Oct 17
- #143 from Oct 18 to Oct 19
- #142 on Oct 20
- #138 from Oct 21 to Oct 22
- #136 from Oct 23 to Oct 24
Before you install
Instructions onlyAutomated signals, not a guarantee
Checked Nov 4, 2026
Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.
Show 47 notes
skills/genoffice/SKILL.md:17Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:18Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:19Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:20Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:21Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:22Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:23Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:24Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:25Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:26Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:27Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:28Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:29Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:30Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:31Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:32Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:33Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:34Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:35Analysis incomplete: physical or joined line exceeds character limitskills/genoffice/SKILL.md:36Analysis incomplete: physical or joined line exceeds character limit
- No broad tool permissions
- No dangerous shell commands
- No network access
- No hidden or obfuscated text
- No instruction overrides
- No executable scripts
- No access to secrets
- No requests to vote in Hot
How the signals fit together
Facts from a static read of the files at the pinned commit. They describe what the code does, not why.
Some files could not be read; the facts below still hold.
No combined signals at this commit.
Works with
| Agent | Status | Evidence | Install to |
|---|---|---|---|
| Claude Code | Native | Path CLAUDE.md | .claude/skills/<skill-name>/ |
| Codex | Compatible | Standard SKILL.md skills/genoffice/SKILL.md | .agents/skills/<skill-name>/ |
| Cursor | Compatible | Standard SKILL.md skills/genoffice/SKILL.md | .cursor/skills/<skill-name>/ |
| Gemini CLI | Compatible | Standard SKILL.md skills/genoffice/SKILL.md | .gemini/skills/<skill-name>/ |
| GitHub Copilot | Compatible | Standard SKILL.md skills/genoffice/SKILL.md | .github/skills/<skill-name>/ |
| OpenCode | Compatible | Standard SKILL.md skills/genoffice/SKILL.md | .opencode/skills/<skill-name>/ |
- Native The repository ships a manifest, folder or frontmatter field for this agent.
- Compatible A standard SKILL.md that this agent can read.
- Unknown No evidence either way.
Derived automatically from paths, manifests and frontmatter. Not tested on each agent.
Add this badge to your README
It updates every night and links to this page.
[](https://hotskills.dev/skill/genspark-ai-genoffice/)