Post-Deployment Validation Agent

AI-powered prototype that validates microservices and web/API services after deployment — checking health, performance, smoke tests, logs, and security.

Requirements

• Python 3.8+ (no extra packages needed — uses stdlib only)
• An Anthropic API key → https://console.anthropic.com

Quick start

# 1. Clone / unzip this folder, then:
cd validation-agent

# 2. Set your API key
export ANTHROPIC_API_KEY=sk-ant-...      # macOS / Linux
# set ANTHROPIC_API_KEY=sk-ant-...       # Windows CMD
# $env:ANTHROPIC_API_KEY="sk-ant-..."   # Windows PowerShell

# 3. Start the server
python server.py

# 4. Open in browser
open http://localhost:8080               # macOS
# xdg-open http://localhost:8080        # Linux
# start http://localhost:8080           # Windows

File structure

validation-agent/
├── server.py          ← Python backend proxy (stdlib only)
└── static/
    └── index.html     ← Full frontend app

How it works

Browser  ──POST /api/claude──▶  server.py  ──▶  api.anthropic.com
         ◀── JSON response ───            ◀──

server.py serves static/index.html on GET requests and proxies POST /api/claude to Anthropic, injecting your API key server-side so it never appears in the browser.

Changing the port

Edit PORT = 8080 at the top of server.py.

Extending with real metrics

Replace the genMetrics() function in index.html with real calls to your monitoring stack (Datadog, Prometheus, CloudWatch, etc.):

async function genMetrics(svcName) {
  const resp = await fetch(`/api/metrics?service=${svcName}`);
  return resp.json();
}

Then add a /api/metrics handler in server.py that queries your actual monitoring APIs.