Add a knowledge source (public website or SharePoint) to a Copilot Studio agent. Use when the user asks to add a knowledge source, documentation URL, website, or SharePoint site for the agent to search.
npx skills add https://github.com/microsoft/skills-for-copilot-studio --skill add-knowledge
Add a knowledge source to the agent. Supports Public Website, SharePoint, and Custom API sources.
Glob: **/agent.mcs.yml
NEVER hardcode an agent name.
sharepoint.com → use SharePointSearchSource (but first normalize the URL — see below)OnKnowledgeRequested (see Custom API section below)PublicSiteSearchSourcehttps://contoso.sharepoint.com/sites/MySite/Shared%20Documents/MyFolder
Users often paste URLs in other formats. Handle them as follows:
| URL pattern | Example | Action |
|---|---|---|
| Direct path (/sites/.../Shared%20Documents/...) | https://x.sharepoint.com/sites/Site/Shared%20Documents/Folder | Use as-is — this is the correct format |
| AllItems.aspx with ?id= param (...AllItems.aspx?id=...) | https://x.sharepoint.com/.../AllItems.aspx?id=%2Fsites%2FSite%2FShared%20Documents%2FFolder&viewid=... | Extract and decode the id query parameter. URL-decode it to get the path (e.g. /sites/Site/Shared%20Documents/Folder), then prepend the origin (https://x.sharepoint.com) to build the direct path. Drop all query parameters (?id=, &viewid=, etc.) |
| Sharing link (/:f:/s/... or /:x:/s/...) | https://x.sharepoint.com/:f:/s/Site/EncodedToken?e=abc | Cannot convert — these are opaque sharing tokens with no extractable folder path. Ask the user: *"That's a SharePoint sharing link — I can't extract the folder path from it. Could you navigate to the folder in SharePoint, copy the URL from the browser address bar, and paste it here?"* |
Encoding rule: Spaces in the final URL must be encoded as %20 (e.g. Shared%20Documents, not Shared Documents).
node ${CLAUDE_SKILL_DIR}/../../scripts/schema-lookup.bundle.js resolve KnowledgeSourceConfiguration
Public Website:
> PublicSiteSearchSource uses Bing search to find relevant snippets within the scoped URL. It does not crawl or summarize full pages. The URL defines a search scope (not a specific page), and supports a maximum depth of 2 levels beyond the domain (e.g. https://example.com/docs/api works, https://example.com/docs/api/v2 is too deep and will be ignored).
# Name: <Name>
# <Description of what this knowledge source provides>
kind: KnowledgeSourceConfiguration
source:
kind: PublicSiteSearchSource
site: https://www.example.com/docs
SharePoint:
# Name: <Name>
# <Description of what this knowledge source provides>
kind: KnowledgeSourceConfiguration
source:
kind: SharePointSearchSource
site: https://contoso.sharepoint.com/sites/MySite/Shared%20Documents/MyFolder
# Name: and a description comment at the top. These are important for identifying the knowledge source.knowledge/<descriptive-name>.knowledge.mcs.yml directoryWhen the user wants to connect a proprietary search API or database not natively supported by Copilot Studio, use the OnKnowledgeRequested trigger. This is a YAML-only capability with no UI designer — it intercepts knowledge search requests and lets you call a custom API to populate results.
How it works:
OnKnowledgeRequested triggerSystem.SearchQuery (the AI-rewritten query) as input to your API callContent, ContentLocation, TitleSystem.SearchResults with the transformed resultsUse the template: ${CLAUDE_SKILL_DIR}/../../templates/topics/custom-knowledge-source.topic.mcs.yml
Prerequisites: The user needs a connector action (in actions/) that calls their search API. If they don't have one yet, they should create it through the Copilot Studio UI or VS Code extension first.
Knowledge source URLs support {VariableName} placeholders — typically global variables — for dynamic routing based on user context. This is powerful for scenarios where different users need different knowledge sources (e.g., geolocation-specific SharePoint folders, region-specific documentation).
Pattern: Define a global variable (via /copilot-studio:add-global-variable), populate it early in the conversation (e.g., in an OnActivity/conversation-init topic based on user profile or geolocation), then reference it in the knowledge source URL with Power Fx string interpolation.
SharePoint example — a workforce agent that routes to different SharePoint folders per region:
# Name: Regional HR KB
# This knowledge source provides HR information from the {Global.UserKBURL} SharePoint.
kind: KnowledgeSourceConfiguration
source:
kind: SharePointSearchSource
site: =$"{Global.UserKBURL}"
Public website example — documentation site with region-specific subpaths:
# Name: Regional Docs
# Documentation for the user's region at {Global.Region}.
kind: KnowledgeSourceConfiguration
source:
kind: PublicSiteSearchSource
site: =$"https://docs.example.com/{Global.Region}/api"
CRITICAL: URL format for the global variable value. The variable must contain a direct, clean URL — not a SharePoint UI URL with query parameters. Copilot Studio cannot resolve AllItems.aspx links.
| Correct (set this in the global variable) | Wrong (will not work) |
|---|---|
| https://contoso.sharepoint.com/sites/MySite/Shared Documents/MyFolder | https://contoso.sharepoint.com/.../AllItems.aspx?id=%2Fsites%2F...&viewid=... |
Note: spaces are OK in the variable value (unlike static site: values which need %20), because the Power Fx $"{Global.UserKBURL}" expression handles encoding at runtime.
Implementation checklist when the user asks for dynamic knowledge:
/copilot-studio:add-global-variable (e.g., UserKBURL, Region)=$"{Global.VariableName}" syntax in the site: fieldFor detailed guidance on how knowledge retrieval works, source selection, content quality, security, maintenance, and advanced patterns like triggerCondition, see knowledge-guide.md.
This skill can create Public Website and SharePoint knowledge sources.
For other types, inform the user:
> "The following knowledge source types must be created through the Copilot Studio UI as they require Power Platform configuration:
> - Dataverse tables
> - Uploaded files
> - AI Search
> - SQL Server
>
> Please create these in the portal, then clone the solution to edit them here."
Integration with protocols.io API for managing scientific protocols. This skill should be used when working with protocols.io to search, create, update, or publish protocols; manage protocol steps and materials; handle discussions and comments; organize workspaces; upload and manage files; or integrate protocols.io functionality into workflows. Applicable for protocol discovery, collaborative protocol development, experiment tracking, lab protocol management, and scientific documentation.
Analyzes job descriptions and generates tailored resumes that highlight relevant experience, skills, and achievements to maximize interview chances
Generate Excalidraw diagrams from natural language descriptions. Use when asked to "create a diagram", "make a flowchart", "visualize a process", "draw a system architecture", "create a mind map", or "generate an Excalidraw file". Supports flowcharts, relationship diagrams, mind maps, and system architecture diagrams. Outputs .excalidraw JSON files that can be opened directly in Excalidraw.
Build and distribute Expo development clients locally or via TestFlight
Use when you have a written implementation plan to execute in a separate session with review checkpoints
Data structure for annotated matrices in single-cell analysis. Use when working with .h5ad files or integrating with the scverse ecosystem. This is the data format skill—for analysis workflows use scanpy; for probabilistic models use scvi-tools; for population-scale queries use cellxgene-census.
Benchling R&D platform integration. Access registry (DNA, proteins), inventory, ELN entries, workflows via API, build Benchling Apps, query Data Warehouse, for lab data management automation.
Comprehensive molecular biology toolkit. Use for sequence manipulation, file parsing (FASTA/GenBank/PDB), phylogenetics, and programmatic NCBI/PubMed access (Bio.Entrez). Best for batch processing, custom bioinformatics pipelines, BLAST automation. For quick lookups use gget; for multi-service integration use bioservices.
Take microsoft/add-knowledge from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.