The gallery of skills for your AI agents

Generate clean, consistently-styled matplotlib charts (bar, line, scatter, histogram, pie) from a DataFrame or CSV with one call.

Detect greenwashing in product descriptions, marketing copy, and catalog entries against EU Directive 2024/825 and the Green Claims Directive. Returns a structured per-claim findings report with risk levels, regulation references, and recommended corrections.

In Copilot Studio, re-sending an edited file under the same name fails to deliver it — the change is made but never reaches the user. This skill gives each iteration a new version-numbered filename (report_v1.docx, report_v2.docx…) so every update actually lands in the chat as its own attachment.

Route Copilot Studio knowledge searches to the right region-specific source (Americas, EMEA, APAC, or Global) based on where the user is, so answers stay locally accurate.

Redlines a document based on changes from a template with Track Changes.

While you're out of office, watches a group chat and answers questions from your local knowledge docs (clearly marked AI-generated), logs anything it can't answer for later, and pings you on Teams if its setup is incomplete.

Convert documents between Markdown, HTML, PDF, Word, PowerPoint, Excel/CSV and text — fully offline, using only the libraries already in the agent sandbox.

Discover, document, and monitor what the agent harness can do — Python libraries, tools, MCP servers, and runtime capabilities — with repeatable, comparable snapshots.

Turns a customer name, line of business, and personas into a ready-to-seed Microsoft 365 Copilot demo — fictional Word/PowerPoint/Excel example files plus a full delivery and provisioning plan, with optional one-click seeding to OneDrive and Teams.

Turn any AI/agentic use case idea (chat, uploaded doc, or attached file) into an evidence-grounded, rubric-scored assessment with a customer-branded HTML report. Guides the user question-by-question through intake, categorisation, strengthening, scoring, and reporting — grounded in the Agentic Use Case Assessment Rubric v2. Runs in Microsoft Scout and Microsoft 365 Copilot Cowork.

Turn engagement notes into anonymised, publishable case studies without exposing client-confidential details.

Audit blog posts or a blog library for clarity, structure, evidence, audience fit, and improvement priority.

Restructure draft or existing blog posts into a stronger narrative without inventing new claims.

Ensure every generated PowerPoint deck or Word document starts from the correct bundled or SharePoint-hosted brand template.

Rewrite drafts into a configurable house style, preserving meaning while removing generic AI phrasing.

Turn a short interview into a charming, vibrant 3-slide marketing-campaign deck by filling a bundled PowerPoint template.

Comments Word or PowerPoint files with Comments.

Schedule a Teams reminder that sends only when the expected person has not already posted a relevant update.

Create CFP-ready talk titles, abstracts, takeaways, speaker notes, and submission copy from a topic.

Score documents, pages, posts, or artefact libraries against a configurable quality rubric and recommend fixes.

A Monday-morning Scout automation that scans authoritative Microsoft sources for the past week's Copilot, Copilot Studio, and agent news and posts a concise, linked digest to Teams.

Reads an exported Copilot Studio agent and generates a graded, runnable test suite (happy-path, paraphrase, disambiguation, negative, knowledge-grounding, multilingual, and safety cases) plus a regression set, ready to run in the free Copilot Studio test panel.

Turn a one-line use case into a build-ready Microsoft Copilot Studio agent blueprint: recommended agent type and orchestration, topics, tools, knowledge, variables, a welcome Adaptive Card, security and Copilot Credits notes, and a first test plan.

Search any Microsoft Dataverse table, browse matching records, and drill into full record details — sourced live via the Dataverse MCP Server.

Build a personalized study plan app for an exam or certification: overview timeline, day-by-day schedule with a calendar picker, domain breakdown, one-click .ics download, and a notes tab to track confidence over time.

Refines an existing plan, decision, or draft through relentless, branch-by-branch questioning. Grouped in batches of 2-5 with recommended answers, pulling facts from Work IQ where possible, until every part has been fully thought through.

When a health system's lab results ingestion pipeline hits an unknown column header, this skill semantically matches the source column against the canonical clinical schema (LOINC-anchored), scoped by lab and clinical domain, and writes a suggested mapping to a Dataverse review queue for a clinical informatics steward to approve.

Create evidence-led LinkedIn posts, promos, newsletter intros, and long-form copy from supplied facts or drafts.

Create credible LinkedIn content from facts, notes or drafts, with configurable voice and no invented evidence.

Analyzes meeting content pasted as text or provided as audio/video transcripts. Delivers a structured intelligence report: explicit decisions and action items, participant persona profiles, and — most importantly — hidden insights: unspoken tensions, implicit risks, unresolved topics, and signals that were not made explicit during the meeting but are critical to the context.

Guides a customer through a discovery interview, recommends the right Microsoft AI platform (M365 Copilot, Agent Builder, Copilot Studio, Foundry, Foundry Agent Service, Windows AI Foundry, or Agent 365 as the governance layer), scores technical complexity and risk, and plots them on a 2x2 quadrant chart with planning guidance.

Extract tables from a PDF document (e.g. contract rebate or pricing schedules) and convert them into a clean, workable Excel spreadsheet or CSV.

Pre-test an internal launch, comms, or enablement artefact against your own role-based personas before it ships — surfaces blackspots, domain gaps, credible-detractor risks, and concrete edits. Bring your own personas file.

Redact the 18 HIPAA Safe Harbor identifiers from clinical text (or produce a Limited Data Set) with consistent pseudonym tokens and an audit manifest of what was removed.

Generate sharp, specific point-of-view lines for posts, slides, talks, campaigns, or positioning work.

Assess Power Automate Desktop and hybrid automation projects with evidence-based findings and prioritized remediation guidance.

Turns a Power Automate solution .zip into a clean markdown reference for every flow inside it: trigger, plain-English process, connection references, and a read/write/delete table for everything it touches. Maps which flows call each other and unresolved connections automatically.

Creates polished PowerPoint decks from a JSON specification using python-pptx. Designed as a Copilot Studio Skill that runs natively inside the agent's Python container (no Azure Function or custom connector required). Supports 8 layouts plus native charts (bar, column, line, pie, donut with stacked / 100% stacked variants).

Writes natural, first-person spoken presenter scripts for the slides in a deck. Treats existing speaker notes as authoritative when they contradict slide text. Calibrates length to a target duration and speaking pace and returns a timing table. Never modifies visible slide content.

Create polished, self-contained HTML slide decks with keynote-style visuals, keyboard navigation, themes, animations, and reusable presentation components.

Turns any SharePoint list data into a clean, consistent markdown table. Dynamic columns based on list type and query, plus a one-click Open link for every row.

Help makers design concise, reusable Agent Skills with clear triggers, instructions, resources, and packaging.

A Scout automation installer that, only during a configured vacation window, scans work email and Teams and forwards genuinely urgent items to your personal email.

Rewrite or draft prose so it avoids the tell-tale signs of AI writing, using Wikipedia's 'Signs of AI writing' field guide.