
Knowledge Work Plugins
https://github.com/anthropics/knowledge-work-plugins- Category
- AI Agents
- Rank
- No. 215Tools index
- Pricing
- Open Source
- Type
- TOOL
- Builder
- anthropics
- GitHub
- 23.9k stars
- Date
About
Anthropic's open-source repository of Claude plugins built for knowledge workers — drop-in skills for research, writing, analysis, and Cowork workflows.
What it does
Knowledge Work Plugins supplies role-specific operating instructions for Claude. Each plugin combines automatic domain guidance, explicit slash-command workflows, external tool connections, and optional organizational context. The collection covers sales, support, product, marketing, legal, finance, data, search, productivity, and biological research, plus plugin management.
Why it's ranked here
The collection has unusual breadth and a clear extension model. Teams can start with a defined professional workflow, then replace connectors and encode their own terminology, processes, playbooks, and risk tolerances. Its value depends heavily on that customization and on access to Claude Cowork or Claude Code.
What's good
The plugins separate automatic expertise from actions users invoke explicitly. MCP connections can pull context from warehouses, CRMs, chat, storage, analytics, and project systems. Several workflows still accept pasted text, uploaded files, or manually supplied results when integrations are unavailable. The Apache license permits modification and redistribution under stated conditions.
Tradeoffs
These are generic starting points, not finished company operating systems. Useful results require teams to add local context and align workflows with internal practice. Some connector entries contain empty URLs and therefore need configuration. Legal guidance requires attorney review, and its supplied examples reflect selected United States jurisdictions until customized.
How to use it well
Pick one role plugin, connect only the systems needed for a recurring workflow, then encode company terms, escalation rules, and preferred outputs. It suits teams already working in Claude Cowork or Claude Code. It does not replace source systems, professional legal judgment, or organization-specific process design.
Technical notes+
README.md describes a marketplace installed through Claude’s plugin CLI, with automatic skills, slash commands, MCP connectors, and file-based customization. data/.mcp.json, legal/.mcp.json, and sales/.mcp.json declare HTTP MCP servers; several entries have blank URLs that require configuration. data/README.md, legal/README.md, and sales/README.md document standalone inputs alongside connector-backed workflows. .github/scripts/external-pr-scope.js implements marketplace contribution scope checks in JavaScript, while bio-research/skills/nextflow-development/scripts/utils/__init__.py exposes Python utilities for NCBI retrieval, file discovery, sample inference, and samplesheet validation. LICENSE, data/LICENSE, legal/LICENSE, sales/LICENSE all contain Apache License 2.0 text.
Observed
- License
- Apache License 2.0
- Platform support
- Designed for Claude Cowork and compatible with Claude Code
- Install surface
- Cowork plugin catalogue and Claude plugin CLI marketplace installation
- Interfaces
- Automatic skills, slash commands, and HTTP MCP connectors
- Content formats
- Plugins are structured primarily as Markdown and JSON configuration
- Supporting languages
- Repository text also includes JavaScript automation and Python bio-research utilities
- Collection structure
- Eleven role or management plugins are listed in the repository documentation
Read from README.md, .github/scripts/external-pr-scope.js, bio-research/skills/nextflow-development/scripts/utils/__init__.py, LICENSE, data/LICENSE, legal/LICENSE, sales/LICENSE, data/.mcp.json, data/README.md, finance/LICENSE, legal/.mcp.json, legal/README.md, sales/.mcp.json, sales/README.md.
What it can do
Conduct research on topics
Research query or topic → Compiled research findings and sources
Generate written content
Writing prompts or requirements → Drafted documents or articles
Analyze data and documents
Data files or text documents → Analysis reports and insights
Execute workflow automation
Workflow specifications → Automated task execution results
Facilitate collaboration tasks
Team collaboration requests → Coordinated team outputs
Process knowledge management tasks
Information organization requests → Structured knowledge repositories
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.