Commands for the daily loop
Create a site, scaffold an app or DocType, run tests with coverage, tail bench logs, manage the Redis cache, back up, migrate and deploy — each one a slash command instead of a half-remembered bench invocation.
AI
Frappe/ERPNext Skills
AI skills for the whole Frappe SDLC
Our AI skills library for the Frappe ecosystem. Autonomous skills cover DocTypes, background jobs, portals, workflows, migrations, microservices and containerisation; specialist agents handle architecture, testing, security and DevOps; commands wrap the everyday bench operations — all grounded in real Frappe and ERPNext core code.
- autonomous skills
- 86autonomous skills
- specialist agents
- 12specialist agents
- commands
- 15commands
- agent platforms
- 3agent platforms
- Language
- AI skills
- License
- MIT
- Agents
- Claude Code · Cursor · Gemini CLI
- Standard
- Agent Skills
- Coverage
- Full SDLC
QUICKSTART
Claude Code
# Two commands, then restart Claude Code/plugin marketplace add vyogotech/frappe-apps-manager/plugin install frappe-apps-manager@vyogotech # Then see what you got/helpUse it
# Scaffold, then build/frappe-new-site/frappe-new-app/frappe-new-doctype # Operate/frappe-test/frappe-logs/frappe-migrate/frappe-deployWhat Frappe/ERPNext Skills gives you
Specialist agents, not one generalist
Twelve of them — architect, developer, UI/UX, tester, security, performance, DevOps, debugger, integration, data engineer, product owner and technical writer — so the advice you get matches the question you asked.
86 skills, loaded on demand
A skill is context the agent pulls in only when the task calls for it, so Frappe-specific depth is available without flooding every prompt with it.
Skills grounded in core code
Skills reference real ERPNext and Frappe core implementations rather than paraphrasing the docs — which is what keeps generated code idiomatic instead of merely plausible.
Microservices and containers included
Skills for microservice scaffolding, tenant isolation, container testing, and generating Containerfiles and compose.yml — wired to the Frappe Microservice Framework.
Modern app standards
Scaffolding targets v14, v15 and v16+ with full pyproject.toml support, module management and modules.txt automation, so a new app starts where current practice is.
One skill set, several agents
Built on the Agent Skills open standard, so the same skills are discoverable by Claude Code, Cursor IDE and Gemini CLI rather than locked to one tool.
Under the hood
Commands, agents and skills are three different tools for three different jobs: a command is a deterministic operation you invoke, an agent is a perspective you delegate to, and a skill is context the model loads when the task calls for it. Splitting them is what keeps Frappe-specific knowledge available without flooding every prompt with it.
Frequently asked questions
How many skills are there?
86 autonomous skills, 12 specialist agents and 15 commands at the time of writing, mirrored across all three supported agent platforms. The set grows — the repository is the authoritative count.
Which agents can use these skills?
The skills follow the Agent Skills open standard, so Claude Code, Cursor IDE and Gemini CLI all read the same set. Each agent discovers them from its own directory; the skill files themselves are shared. Installation into Claude Code goes through its plugin marketplace.
What is the difference between a command, an agent and a skill?
A command is something you invoke deliberately — /frappe-new-doctype does one job. An agent is a specialist you hand a problem to, like security review or performance work. A skill is knowledge the model pulls in automatically when the task needs it, such as how Frappe background jobs or portal pages actually work.
Does it cover microservices?
Yes. There are skills for microservice scaffolding, tenant isolation and container testing, plus automatic Containerfile and compose.yml generation during app creation, integrated with the Frappe Microservice Framework.
Which Frappe versions does the scaffolding target?
v14, v15 and v16+, with pyproject.toml support and modules.txt automation, so a newly scaffolded app matches current Frappe packaging conventions rather than the older setup.py layout.
Works well with
MCP Server
AI bridge for ERPNext
A Model Context Protocol server that lets Claude, Cursor and any LLM query ERPNext safely.
ExploreFrappe AI
AI in the Awesome Bar
Turns Frappe's global search into a natural-language assistant that respects permissions.
ExploreMicroservices Framework
Domain-driven Frappe services
A Python framework for carving secure, multi-tenant services out of the Frappe monolith.
ExploreRunning Frappe/ERPNext Skills in production?
We build and maintain these tools, and we help teams adopt them. Tell us what you are trying to ship and we will tell you honestly whether this is the right piece.
Talk to the team