Knowledge Work Plugins

Knowledge Work Plugins is an open-source collection of Claude plugins published by Anthropic on GitHub under the Apache License 2.0. Built primarily for Claude Cowork and also compatible with Claude Code, each plugin bundles domain expertise, slash commands, and MCP-based tool connectors for a specific job function. The repository was created in January 2026 and, according to GitHub data, has accumulated over 16,000 stars and nearly 2,000 forks.

What It Is

Knowledge Work Plugins is a plugin marketplace and framework that turns Claude into a role-specific specialist. Rather than a single general-purpose assistant, each plugin encodes the workflows, terminology, and tool integrations relevant to a particular function — sales, finance, legal, data analysis, customer support, and others. The plugins are file-based (markdown and JSON), require no code, no infrastructure, and no build steps, making them accessible to non-engineers who want to customize Claude's behavior for their team.

Plugin Catalog and Coverage

The repository ships 11 plugins at launch:

• productivity — task management, calendars, and daily workflows; connects to Slack, Notion, Asana, Linear, Jira, Monday, ClickUp, and Microsoft 365
• sales — prospect research, call prep, pipeline review, outreach drafting, and competitive battlecards; connects to HubSpot, Close, Clay, ZoomInfo, Fireflies, and more
• customer-support — ticket triage, response drafting, escalation packaging, and knowledge base article creation; connects to Intercom, HubSpot, Guru, and Jira
• product-management — spec writing, roadmap planning, user research synthesis; connects to Figma, Amplitude, Pendo, and Linear
• marketing — content drafting, campaign planning, brand voice enforcement; connects to Canva, Ahrefs, SimilarWeb, Klaviyo, and HubSpot
• legal — contract review, NDA triage, compliance navigation, and risk assessment; connects to Box, Egnyte, and Microsoft 365
• finance — journal entries, reconciliation, financial statements, variance analysis; connects to Snowflake, Databricks, and BigQuery
• data — SQL writing, statistical analysis, dashboard building, and data validation; connects to Snowflake, Databricks, BigQuery, Hex, and Amplitude
• enterprise-search — unified search across email, chat, docs, and wikis
• bio-research — preclinical research tools including PubMed, BioRender, ChEMBL, ClinicalTrials.gov, and Benchling
• cowork-plugin-management — a meta-plugin for creating and customizing other plugins

How Plugins Are Structured

Every plugin follows a consistent directory layout: a plugin.json manifest, an .mcp.json file for tool connections, a commands/ folder for explicit slash commands, and a skills/ folder for domain knowledge Claude draws on automatically. Skills encode best practices and step-by-step workflows; commands are explicit actions users trigger (e.g., /finance:reconciliation, /sales:call-prep); connectors wire Claude to external tools via MCP servers.

Customization Model

The plugins are designed as generic starting points. Teams can swap connectors by editing .mcp.json, add company-specific terminology and org structure to skill files, adjust workflow instructions to match internal processes, and build entirely new plugins using the same structure. The cowork-plugin-management plugin provides a guided path for creating new plugins without leaving the Claude interface.

Deployment and Setup

Plugins can be installed directly from claude.com/plugins for Cowork users. For Claude Code, the repository can be added as a marketplace source and individual plugins installed via the claude plugin install CLI command. Once installed, plugins activate automatically — skills fire when contextually relevant, and slash commands become available in the session.

Update: Initial Open-Source Release

The repository was created on January 23, 2026, and last pushed on May 26, 2026, indicating active development. The project is licensed under Apache 2.0, making it freely forkable and modifiable. With over 16,000 GitHub stars and nearly 2,000 forks reported at the time of data collection, the repository has attracted significant community interest since its launch.