惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

U
Unit 42
Vercel News
Vercel News
博客园 - 叶小钗
大猫的无限游戏
大猫的无限游戏
MyScale Blog
MyScale Blog
P
Proofpoint News Feed
量子位
Engineering at Meta
Engineering at Meta
B
Blog RSS Feed
博客园 - 【当耐特】
Recent Announcements
Recent Announcements
Google DeepMind News
Google DeepMind News
D
DataBreaches.Net
Stack Overflow Blog
Stack Overflow Blog
博客园 - 聂微东
小众软件
小众软件
Hugging Face - Blog
Hugging Face - Blog
人人都是产品经理
人人都是产品经理
IT之家
IT之家
T
The Blog of Author Tim Ferriss
Last Week in AI
Last Week in AI
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
Jina AI
Jina AI
博客园 - 三生石上(FineUI控件)

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant
I build an AI agent using StackOne MCP
prathyusha k · 2026-05-06 · via DEV Community

prathyusha k

Hello myself Prathyusha. When I decided to apply to StackOne, I did not send
a resume first. I built something with their platform first.

This is the story of building an AI agent using StackOne MCP.

What I Built

An AI agent that onboards a new employee across
4 enterprise platforms automatically — using
StackOne MCP as the single integration gateway.

One command does everything:

npm run onboard --
example employee name: Priya Sharma

Output:

Step 1: Fetching candidate... ✓ Priya Sharma
Step 2: Creating BambooHR record... ✓ Employee #115
Step 3: Creating Notion page... ✓ Checklist created
Step 4: Sending Slack message... ✓ Team notified

Onboarding complete in seconds.
Audit trail written to audit-log.json.

The Problem It Solves

Most companies onboard employees manually:

  • HR types data into BambooHR by hand.
  • Someone creates a Notion doc from scratch.
  • Manager writes a Slack welcome manually.
  • IT gets notified separately.

That is 2 to 4 hours of manual work per new hire.
This agent does all of it in seconds.

What is StackOne MCP?

  • MCP stands for Model Context Protocol.
  • It is an open standard that lets AI agents
  • connect to external tools in a structured way.

Without StackOne:
Agent → custom code → BambooHR
Agent → different code → Notion
Agent → another code → Slack

With StackOne MCP:
Agent → ONE endpoint → BambooHR
→ Notion
→ Slack

One API key. One endpoint. 260+ integrations.
Auth, retries, and audit logging handled
automatically by StackOne's Falcon engine.

The Architecture

CLI input

src/index.ts — reads candidate name

src/agent.ts — orchestrates 4 steps

src/stackone.ts — MCP client

api.stackone.com/mcp
↓ ↓ ↓
BambooHR Notion Slack

audit-log.json

Built in TypeScript. 9 files. Clean structure.
Supports multiple candidates in one command:

npm run onboard -- john sarah alex priya

Tech Stack

Language → TypeScript
Runtime → Node.js + tsx
Gateway → StackOne MCP server
HRIS → BambooHR
Docs → Notion
Messaging → Slack
Audit → JSON file

3 Real Errors I Hit

These taught me the most about StackOne MCP.

─────────────────────────────────────────
Error 1 — Missing Accept Header

Not Acceptable: Client must accept both
application/json and text/event-stream

Why: StackOne MCP uses streaming protocol.
Must declare support for both formats.

Fix: Add this header to every request:

"Accept": "application/json, text/event-stream"
─────────────────────────────────────────
Error 2 — Silent BambooHR Redirect

No error thrown. Agent said success.
Employee never appeared in BambooHR.

The response was returning:
"location": "https://www.bamboohr.com"

Instead of:
"location": "https://api.bamboohr.com/
.../mytestorg/.../employees/115"

Why: Wrong BambooHR account connected
in StackOne. Requests went to wrong instance.

Fix: Delete wrong account in StackOne.
Re-link with correct subdomain: mytestorg.
─────────────────────────────────────────
Error 3 — Wrong Slack Tool Name

MCP error -32602: Tool slack_post_message
not found

Why: StackOne names tools differently
from the underlying provider APIs.

Fix: Called tools/list to see all Slack tools.
Correct name was slack_send_message.
Field name was channel_id not channel.

Pro tip: Always use tools/list before
assuming tool names. Saves hours of guessing.
─────────────────────────────────────────
What I Learned About StackOne MCP

  • Always accept text/event-stream in headers.
  • Each integration has its own Account ID.
  • Use tools/list to find exact tool names.
  • All tools wrap fields inside body object.
  • Wrong account = silent redirect, not error.
  • StackOne Logs shows every API call made.

Results

BambooHR → Employee #115 created ✓
Notion → Onboarding page with Day 1 checklist ✓
Slack → Welcome message in #onboarding-alerts ✓
Audit → 4 timestamped entries logged ✓

Final Thought

StackOne MCP made connecting to 4 enterprise
platforms easier than connecting to one platform
the traditional way.

One API key replaced four SDKs.
One endpoint replaced four auth systems.
One integration replaced months of work.

Full code on GitHub:
github.com/prathyushak828/stackone-onboarding-agent

If you are building AI agents that need to touch
enterprise SaaS — StackOne MCP is worth your time.

StackOne #MCP #TypeScript #AIAgents #BuildInPublic