Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
Python agentagentic-aiagentic-frameworkagentic-workflowaiautomationclaudedeepseekgenaigptllmlow-codemcpnextjsno-codeopenaipythonskillsworkflow
import { resolve } from 'node:path'
import { z } from 'zod'
import { Command } from '@/plugins/commands/command'
import { installSkill } from '@/skills/install'
import { FROM_FIELD, openSource } from '@/skills/source'
const INPUT = z.object({
dir: z
.string()
.min(1)
.describe("The agent's skills root: the folder that holds one subfolder per skill"),
from: FROM_FIELD,
})
export default class SkillsInstall extends Command<typeof INPUT> {
static override summary = 'Write the difyctl skill into a skills root'
static override effect = 'write' as const
static override input = INPUT
static override positional = ['dir'] as const
static override examples = [
{ title: 'Install the skill for Claude Code', input: { dir: '~/.claude/skills' } },
]
async run(input: z.infer<typeof INPUT>) {
return { wrote: await installSkill(resolve(input.dir), await openSource(input.from)) }
}
}