Skip to content

AgentlyHQ/workflows

Repository files navigation

Agentic Workflows

Multi-agent workflows for use-agently.com.

Each workflow is a SKILL.md file that orchestrates multiple AI agents from the Agently marketplace into a single, outcome-focused job. Instead of calling agents one at a time, workflows chain them together — search, extract, synthesize, render — so you get results, not raw API calls.

Workflows are designed to be run by AI coding agents (Claude Code, Cursor, etc.) using the use-agently CLI as the gateway for all agent interactions.

Install as Skills

Each workflow can be installed directly as a Claude Code skill:

npx skills add https://github.com/agentlyhq/workflows

Available Workflows

Workflow Category Description
AI-Native Team Research Research Multi-source intelligence sweep on how AI-native teams operate in any domain
AI Search Toolkit Research Deep research across multiple search engines and domain-specific sources
Market Intelligence Brief Research Produces a structured, multi-source market intelligence brief for any topic with investment-relevant insights, sector dynamics, company signals, and actionable watchpoints
X.com Marketing Pulse Marketing Pull the most recent X.com tweets on a topic and produce a deterministic, engagement-ready summary plus a draft reply
GEO Site Audit Marketing Pre-checks a target site with Jina, then delivers a weighted GEO audit with scores and a prioritized action plan
Is It Worth Building? Research Ruthlessly evaluates your idea — researches market demand, competition, timing, and feasibility, then delivers an honest verdict on whether you should build it
Singapore HDB Resale Analysis Research Buy/don’t-buy assessment for a Singapore HDB area or block using recent resale comps, trends, lease profile, and caveats
Security Toolkit Security Multi-layered security audit: secret scanning, dependency vulnerabilities, MCP server trust, AI prompt injection testing, and penetration testing
Singapore Condo Analysis Research Converts pasted URA condo transaction data plus local context into a grounded buy / wait / avoid recommendation with 4-year post-SSD scenarios

API

Served at workflows.use-agently.com:

  • GET /api/workflows — list all workflows
  • GET /api/workflows/:slug — get a single workflow with full prompt content

SKILL.md Format

Each workflow lives in workflows/<slug>/SKILL.md with YAML frontmatter:

---
name: AI-Native Team Research
description: Multi-source intelligence sweep on how bleeding-edge AI-native teams operate...
license: MIT
metadata:
  category: Research
---

(prompt content referencing agents and variables)
  • Agents and variables are auto-extracted from the prompt content
  • Agent references link to use-agently.com marketplace agents via [Name](https://use-agently.com/agents/<erc-8004-uri>) links
  • Template variables use ${UPPER_SNAKE_CASE} syntax — the consumer app auto-detects these and prompts for input

Development

bun install
bun run dev

About

Workflows are pre-built, executable task chains that combine agents, tools, and data sources into ready-to-run, repeatable jobs — turning "how to do this" into "this gets done."

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors