Why this matters
Most enterprise AI deployments fail at the final mile: translating team processes, tooling, and terminology into prompts that reliably produce useful outputs. This repository treats that translation as code — markdown- and JSON-based “plugins” that package domain skills, explicit slash commands, and MCP connector metadata so Claude can behave like a specialist for a role or team without bespoke engineering for each workflow.
What Sets It Apart
- File-first, opinionated packaging: every plugin is just markdown and JSON (manifest, .mcp.json, skills, and commands), which makes customsation, review, and versioning simple — you edit files, open a PR, and the behavior changes.
- Role-focused building blocks: the repo ships a curated set of plugins (productivity, sales, customer support, product management, marketing, legal, finance, data, enterprise search, bio-research, and plugin-management) that combine procedural skills with connectors to common enterprise tools.
- Protocol-native integrations: connectors are expressed for use with MCP servers, so the plugins are meant to be wired into existing tool auth and data sources rather than reimplementing integrations inside the model prompt.
- Designed for Claude ecosystem but reusable: plugins work out-of-the-box in Claude Cowork and Claude Code; because they're file-based they can be adapted to other MCP-compatible runtimes with engineering effort.
Who It's For and Trade-offs
Great fit if you manage knowledge workers and want predictable, shareable LLM behaviors — product, sales, support, legal, and data teams can convert playbooks, templates, and connectors into reproducible assistant skills.
Look elsewhere if you need a standalone LLM app: these plugins are not a complete UI or hosting layer and expect an MCP-compatible runtime (Claude Cowork/CODE or an MCP server). Adoption requires connector configuration and attention to data access and compliance (the plugins reference external tools and thus rely on secure credentialing and enterprise governance).
Where It Fits
Use this as a starting marketplace for organizational customization: fork or extend the plugins to add company terminology, swap connectors to your tool stack, and bake operational best practices into the assistant so team members get consistent outputs instead of ad-hoc prompt engineering.
