starttoscale

Draft, preview, and propose reviewed website or brand edits through the asset-broker MCP without direct GitHub access. Use when a client wants to add or edit their own website content (blog, case studies, intel, capabilities, images) or brand/charter data and have it land as a reviewed GitHub pull request with a truthful preview.

Capture structured client, operator, and agent-execution feedback through the asset-broker MCP and write it to the scoped feedback lane in the shared `client-feedback` repo. Use when the user wants to record product feedback on a deliverable or when the agent should file an infra-quality retrospective about the skill/tooling it just used.

Product guide for editing your brand and website through the asset-broker connector — what each skill does, how a draft becomes a truthful preview and a reviewed pull request, and how to give feedback. Use whenever someone asks "how do I edit my brand/website/logo/colours", "how do I give feedback", "how do I change my site content", or any orientation question specifically about brand or website edits. For first-time install, switching on connectors, or a general "how do I get started", the `getting-started` skill is the entry point.

Design or change how a client's own SharePoint workspace is organised — the folder tree deliverables land in, what its stages are called, and whether the space keeps a visible project record. Interviews the client about how they actually work, then proposes the configuration as a reviewed pull request through the asset-broker MCP. Use when setting up a new client's or partner's workspace, when someone asks where their deliverables go or why something landed in the wrong place, or when they want to rename or restructure their folders.

Render brand-aware data-visualisation charts using Plotly.js. The LLM authors any Plotly figure JSON (30+ chart types — waterfall, sankey, sunburst, treemap, funnel, parallel coords, calendar heatmap, candlestick, gauge, radar, themeRiver, marimekko-via-stacked, etc.) and this skill applies the client's brand theme from charter.plotly + tokens.css, then renders to PNG/SVG for embedding in PPTX/DOCX/PDF. Use whenever the user asks for a chart, data visualisation, plot, graph, KPI viz, or asks to add a quantitative visual to a deliverable. Defers to the `format-diagram` skill for structural visuals (process flows, org charts, stakeholder maps) — `chart` is purely numerical data-viz.

Generate branded diagrams (process flows, architecture views, stakeholder maps, org charts, timelines, funnels, matrices, mind maps) as SVG/PNG for embedding in deliverables. TWO routes: (1) the DEFAULT for flow/relationship diagrams — author Mermaid text and call the server-side `render_diagram` MCP tool (auto-layout + deep brand theming, works in a deployed sandbox, brand-gated); (2) hand-authored Excalidraw JSON for bespoke structural art (operator environment only). Use when asked to create a diagram, draw a process flow, make an architecture diagram, visualize a workflow, create a stakeholder map, org chart, funnel diagram, timeline, or any structural visual.

Create, read, edit, and manipulate Word documents (.docx files). Triggers on any mention of 'Word doc', 'word document', '.docx', or requests to produce documents with formatting like tables of contents, headings, page numbers, headers/footers, or letterheads. Also use when extracting or reorganizing content from .docx files, inserting or replacing images, performing find-and-replace, working with tracked changes or comments, or converting content into a polished Word document. If the user asks for a 'report', 'memo', 'letter', 'brief', or 'template' as a Word file, use this skill. When a single client overlay resolves, use it automatically for branding unless the user explicitly asks for unbranded output. Do NOT use for PDFs (use the `format-pdf` skill), spreadsheets (use `format-xlsx`), Google Docs, or coding tasks unrelated to document generation.

Product guide for producing documents, decks, and other deliverables — what to ask for, which format to pick, how the skills hand off to each other, and where finished work lands. Use whenever someone asks "how do I make a document/deck/PDF", "how do I create a proposal/presentation", "which format should I use", or any orientation question specifically about producing a deliverable. For first-time install, switching on connectors, or a general "how do I get started", the `getting-started` skill is the entry point. Routes to the right renderer; does not render anything itself.

Author a premium, fully self-contained branded HTML slide deck — ONE .html the agent writes directly on the client (no server render, no reveal.js): container-query 16:9 stages, scroll-snap + keyboard nav, inline-SVG brand motifs & diagrams, and the client's real fonts/logos base64-inlined so it opens offline and shares as a single file. Deeply brand-integrated from the invoking plugin's companies/<slug>/ overlay — a deterministic design system keeps every deck unmistakably THIS client's, while a seeded variance engine makes each deck fresh. USE THIS whenever someone wants an HTML / web / browser slide deck, a shareable single-file presentation, an interactive deck, a good-looking or on-brand web deck, or to turn a topic, brief, or report into branded slides — even if they don't say 'HTML'. Siblings: editable PowerPoint → format-pptx-hd; paginated print PDF → format-pdf-hd; the older server-rendered reveal.js deck → format-html-reveal.

Create a premium, self-contained reveal.js HTML slide deck — ONE downloadable .html (reveal.js + brand fonts + images all inlined) that opens offline in any browser and shares as a file or link, SERVER-RENDERED via the `render_deck` MCP tool (brand gate, no local build). Premium-floor layout archetypes + a deterministic variance engine (brand_context.expression.visualAxes + surfaceExpression.presentation) keep every deck premium AND visibly each client's own. TRIGGER on: 'reveal.js deck', 'server-rendered web deck', 'interactive presentation', 'shareable web slides'. SIBLINGS — pick the right one: a hand-authored client-side self-contained HTML deck (no reveal.js, full pixel control) → format-html-hd; an editable .pptx → format-pptx-hd; a paginated print PDF → format-pdf-hd.

Maintain a deliverable that exists in several languages without the versions drifting apart. Reports exactly which translation units are missing, stale, current or obsolete for a target locale; returns only the units that must be translated; validates a translated batch against the source's protected tokens, placeholders, URLs, figures and terminology; and hands back a replacement target state for the caller to persist. Use when asked to translate a deliverable, add a language version, check whether a translation is up to date, find out what changed since the last translation, or work out why an English copy still has source-language text in it. Triggers on "translate this", "add an English version", "is the translation current", "what changed since we translated it", "the translation is out of date", "half of it is still in Dutch", "keep the language versions in sync".

Creates Mermaid diagrams for Markdown-based documentation (GitHub, GitLab, wikis, blogs) and visual explanations. Supports 25+ diagram types including flowcharts, sequence/class/state diagrams, ERDs, Gantt charts, mindmaps, timelines, architecture views, Kanban boards, Sankey diagrams, and newer types (Wardley maps, Venn, Ishikawa). When the `mermaid` MCP is available, uses live browser preview and PNG/SVG/PDF export. Use this skill when users ask for Mermaid syntax, Markdown-renderable diagrams, visual architecture, or quick text-to-diagram output.

Author tasteful, brand-aware motion across Stromy''s HTML surfaces — a substrate-agnostic library of motion primitives (staggered reveal, line-draw, count-up, motion-path, scroll-reveal, element morph) plus the restraint doctrine that keeps motion meaningful. Two modes: embed-live (inline a primitive into a deck or site you are hand-authoring) and bake (author a self-contained HTML animation and turn it into MP4/GIF/PNG frames via the render_motion MCP tool, for PowerPoint or PDF). USE THIS whenever someone wants animation, an animated reveal or transition, a line that draws itself, a counting KPI, an animated explainer, a baked clip from web animation, or a deck or site that feels alive without becoming gimmicky. Brand-optional: reads the brand''s motion grammar from companies/<slug>/brand_context.json when present. Siblings: numeric data-viz → format-chart; structural diagrams → format-diagram; full branded explainer video → format-video-hd.

Comprehensive PDF manipulation toolkit for extracting text and tables, creating new PDFs, merging/splitting documents, and handling forms. When Claude needs to fill in a PDF form or programmatically process, generate, or analyze PDF documents at scale.

High-fidelity branded PDF creation using HTML-first design with the full web stack (CSS gradients, web fonts, SVG, paged-media CSS). Server-renders via the `render_pdf` MCP tool (Playwright/Chromium + brand gate) — no local build script. Deeply integrated with the invoking plugin's brand overlay (`brand_context.json`, `assets.json`, hero images). Use when asked to create branded proposals, executive briefs, brand books, policy reports, case studies, white papers, or any client-facing PDF where visual quality matters. Triggers on: 'create branded PDF', 'build proposal PDF', 'design a brief', 'brand book PDF', 'high quality PDF', 'HD PDF', 'magazine-style PDF', or any request for visually polished branded paginated output.

DEFAULT PPTX skill for **unbranded deck creation** and **editing / analysis / OOXML / comments / speaker-notes of any existing deck** (branded or not — in-place OOXML edits, not a branded render). For a **net-new branded deck**, use `format-pptx-hd` (server-rendered + brand-gated). This skill does not render a branded deck from scratch. Also supports `render-anchors` mode (materialises brand anchor templates from spec files).

TRIGGER on **net-new branded deck creation** (a client charter exists) — pitch/investor/executive **and** ordinary branded new decks — plus explicit HD cues and `render-anchors`. Server-renders via `render_pptx` in the brand fonts with a brand gate, web-openable by default (no local build-script path). HTML-first design with full web stack. Deeply integrated with the invoking plugin's brand overlay (`charter.json`, `tokens.css`, `assets.json`, hero images, boilerplate.json, anchor templates in `templates/pptx/`). **Editing or analyzing an existing deck (branded or not) stays on `format-pptx`** — `pptx-hd` regenerates from HTML and cannot edit in place. Unbranded quick decks also go to `format-pptx`.

Plan a multi-section document before rendering it. Use when a branded PPTX, DOCX, or PDF needs structure-first collaboration: section plan, substrate choice, complementary charts/diagrams, sign-off, then handoff to the correct `format-*` renderer. This skill is generic and client-agnostic apart from the invoking plugin's brand overlay.

Author a premium branded explainer VIDEO — a server-rendered MP4 with animated titles, kinetic typography, animated process-flow diagrams, animated data (data-bars, KPI count-ups), image reveals, captions, and optional voiceover or music. Storyboard-first and accuracy-gated: plans a scene-by-scene shot list with per-scene citations, runs the voice cascade, then renders via the async `render_video` MCP tool behind a brand gate and a per-scene visual review. Brand-required, wired to the plugin''s overlay (`brand_context.json`, `assets.json`). USE THIS whenever someone wants a video, explainer, animated explainer, motion-graphics video, product/strategy/plan walkthrough, or a brief, plan or report turned into a branded video — even if they never say ''video''. Content is condensed and accurate, never fabricated. Siblings: an HTML-to-MP4 clip for a deck or PDF → format-motion; generative AI footage → video-production; an editable branded deck → format-pptx-hd.

Load and maintain the project record in a client's SharePoint workspace — the shared, client-readable account of what a project is, what was decided, what is open, and what has been delivered. Use at the START of work on an existing project (so you continue instead of re-briefing), and again whenever a meaningful milestone, decision, client correction, risk, or delivery happens. Triggers on "what's the status of this project", "pick up where we left off", "continue the campaign", "record that we agreed X", "log this decision", "update the project record", or any request to resume or capture the state of ongoing client work.

Comprehensive spreadsheet creation, editing, and analysis with support for formulas, formatting, data analysis, and visualization. When Claude needs to work with spreadsheets (.xlsx, .xlsm, .csv, .tsv, etc) for: (1) Creating new spreadsheets with formulas and formatting, (2) Reading or analyzing data, (3) Modify existing spreadsheets while preserving formulas, (4) Data analysis and visualization in spreadsheets, or (5) Recalculating formulas

Day-0 setup and orientation for a Stromy plugin — install it, switch on the connectors it needs, confirm it worked, and find which guide answers your question. Use whenever someone is new, asks how to get started, how to install or set up the plugin, what they have available, or hits one of the two common snags: skills that look out of date after an update, or a tool that reads "Connected" but will not run in the chat. This skill ships inside the plugin and needs no connector, so it is the one that still answers when nothing else is switched on.

Build structured messaging frameworks — core narrative, messaging pillars, proof points, audience adaptations, and verbal guardrails — from positioning and research inputs. Supports five framework types: Message House, Messaging Hierarchy, Messaging Matrix, Strategic Narrative, and Product Messaging. Produces format-agnostic messaging architecture that downstream skills (talking-points, media-pitch, campaign briefs) can consume. Integrates with company profiles for branded output and maintains a reusable messaging content library. Use this skill whenever the user asks to build a messaging framework, create a message house, develop messaging pillars, write a messaging hierarchy, build a messaging matrix, define key messages, create a messaging architecture, develop brand messaging, write positioning messages, structure audience-specific messaging, or anything involving 'what should we say and to whom' — even if they just say 'we need messaging for X' or 'help me figure out what to say about this.'

Write and manage corporate press releases with full governance lifecycle — from newsworthiness assessment through drafting, legal review, and distribution planning. Produces AP-style, journalist-ready releases with inverted pyramid structure, proper datelines, sourced quotes, and boilerplate from company data. Covers all announcement types: product launches, funding rounds, partnerships, acquisitions, executive hires, milestones, research/reports, and crisis communications. Integrates with company profiles for branded output, spokesperson management, and approval workflows. Use this skill whenever the user asks to write a press release, draft a media announcement, create a news release, prepare a PR statement, write an announcement for the press, handle media communications, or produce journalist-ready content — even if they just say 'announce this' or 'we need to tell the media about X'.

Build production-quality consulting proposals, executive briefs, and bid documents. Supports 5 proposal archetypes (Corporate, Consultant, Entrepreneur, Creative, Minimal) with quick-setup and full intake workflows. Format-agnostic — handles proposal strategy, content assembly, and quality validation, then produces the final document in the requested output format (DOCX, PPTX, PDF). Use when creating new proposals, revising existing ones, merging sections, or condensing documents.